担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

In Japan, the population of working age between 15 and 64 years old peaked in 1995 at about 87 million and is expected to continue to decline in the future. Therefore, to solve the labor shortage, the introduction of industrial robots that can perform the same level of work as humans is strongly requested especially in the food preparation industry. In order to prevent danger to workers there, industrial robots must recognize workers and avoid them when there is fear of collision. In this paper, we propose a method for extracting hand regions based on the color distribution of a hand and GrabCut in an experimental environment to recognize human hands and detect their directions of approach. The proposed method was examined experimentally and gave satisfactory results.

Kyutacar

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper describes a method of navigating a visually impaired person who walks on a sidewalk and goes through a pedestrian crossing by the use of an ego-camera system named MY VISION. The proposed method finds a walkable region on the sidewalk. Thus, at a pedestrian crossing, it finds a traffic light and judges an appropriate timing to cross the road by analyzing the color and flashing status of the signal. The walkable region is found employing Graph- Based Segmentation (GBS), and once a pedestrian crossing is reached through the use of a GPS signal, a traffic light is detected by use of the HOG feature and Random Forest. The effectiveness of the proposed method was verified by experiments. This implies usefulness of a MY VISION system.

Kyutacar

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

In response to requests from visually impaired people for better assistance tools in their daily lives and due to the difficulty of using public transportation, it is proposed a bus boarding support system using MY VISION. For the support, a method of identifying a bus route number is proposed in this paper. A bus approaching a bus stop is detected from MY VISION images that utilizes the Lucas-Kanade tracker. A bus route number area is then acquired with the help of the Haar-like filters and the random forest. Finally, the number from the area is extracted and identified through pattern matching. The effectiveness of the proposed method was shown by experiments. The results indicate that it is possible to realize a practical boarding support system for the visually impaired.

DOI： 10.2991/jrnal.k.210922.014

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85120943766&origin=inward

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper proposes a method of detecting a person fallen on the ground and its head and lower body from aerial images. The study intends to automate discovering victims of disasters such as earthquakes from areal images taken by a UAV. Rotation-invariant HOG and rotation-invariant LBP are employed as features describing a fallen person so as to detect it regardless of its body orientation. The proposed method also detects the head and the lower body of a fallen person using the peak of the gradient histogram. Experimental results show satisfactory performance of the proposed method.

DOI： 10.2991/jrnal.k.210713.013

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85120707754&origin=inward

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

A system is proposed that performs gait analysis of a pedestrian to improve a walk posture and at the same time to prevent fall. In the system, a user walks with a chest-mounted camera. His/her walking posture is estimated using a pair of images obtained from the camera. Normally it is difficult to estimate the camera movement, when the parallax of the image pair is small. Therefore, the system uses a convolutional neural network. Optical flow and camera movement, and depth images are estimated alternately. Satisfactory results were obtained experimentally.

Keywords: Posture, Posture analysis, Pedestrian, MY VISION, Depth and Motion Network.

DOI： 10.2991/jrnal.k.200909.002

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85097097973&origin=inward

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper proposes a human-robot cooperative system supported by visual communication. Human-robot cooperation is expected in various fields nowadays in order to raise efficiency, accuracy and safety of work, to name a few, production of goods, logistics, car driving, care of aged people, etc. In the human-robot cooperation, it is desirable that a human takes less part of work; whereas a robot takes a larger role with it. The idea of the present paper is that visual decision making is done by a human rather than a robot. The proposed system goes to a specified remote spot autonomously and performs objects acquisition there by communicating visually with a user. It aims at realizing ‘a shopping robot’ in near future and expected to be used by disadvantaged shoppers. The proposed system is presented and some experimental results are shown.

DOI： 10.24507/ijicic.16.02.543

Kyutacar

Scopus

CiNii Article

CiNii Research

その他リンク： https://www.scopus.com/record/display.uri?eid=2-s2.0-85082019101&origin=resultslist&sort=plf-f

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

In recent years, research on computer vision has shown great advancement and has been applied to a wide range of fields. Among them, automatic recognition of human motion is an important technology especially in crime prevention and elderly watching systems. Considering this trend, the paper proposes a novel method of human motion description and recognition employing a MHI (Motion History Image)-based method called TMRIs (Triplet Motion Representation Images) and an extended feature descriptor called HOOF (Histograms of Oriented Optical Flow) which contains information on the direction and velocity of movement. One of the advantages of the proposed method over existent methods is that it solves a self-occlusive motion problem particularly in the depth direction which occurs when a single camera is used. The performance and effectiveness of the proposed method are verified by experiments.

Keywords: Human motion; Description; Recognition; Elderly care; crime prevention; MHI; Triplet Motion Representation Images; HOOF.

DOI： 10.2991/jrnal.k.201215.004

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85103065635&origin=inward

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper proposes a pedestrian detection method using a MY VISION system. The MY VISION system is an image processing system employing an ego-camera which a user of the system possesses, and it is expected to be the third eye of those who are aged, visually impaired or even those who are absorbed in a mobile-phone while walking. In the proposed method, a flow area different from camera movement is extracted first on the image provided from a user’s head-mounted camera. The area is a candidate area where pedestrians may exist. MSC-HOG features are then calculated to detect pedestrians in the area. HOF feature is further computed to recognize the moving direction of the detected pedestrians. The proposed method was examined its performance experimentally and satisfactory results were obtained.

Keywords: MY VISION, Ego-cameras, Optical flow, MSC-HOG, HOF, Pedestrian detection, Walk direction.

DOI： 10.2991/jrnal.k.200528.014

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85097079744&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper proposes an autonomous mobile robot expected to realize an intelligent robot that supports a human life.The proposed mobile robot has three main functions; self-position estimation, road region estimation, and route planning. Self-position estimation is performed by comparing local features obtained from the frontal images captured
by the robot to a knowledge base. In road region estimation, the frontal image is separated into several regions to find　the region where the robot can move. Furthermore, the route to the destination is planned by graph search.
Experimental results show satisfactory performance of the proposed mobile robot.

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2024年02月22日  -  2024年02月25日

It is crucial for individuals to keep walking and stay healthy to prevent receiving nursing care. This paper proposes a method of recognizing walk motions and analyzing the gait cycle of a human focusing on his/her posture. We use 43 structural features defined from human joint coordinates obtained using OpenPose and 18 figural features from human domain images and their difference images. The feature vector containing these 61 features is used for the recognition of walk motion by Random Forest. In the experiment, we applied the method to recognizing six types of motions and analyzed the walk gait cycles of five persons, and obtained satisfactory results.

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2024年02月22日  -  2024年02月25日

In recent years, elderly people living alone account for a large proportion of the elderly population, and the issue of safety has also been a matter of great concern for the public. Considering the importance of monitoring the behavior and activities of the elderly and detecting abnormal movements, this paper proposes a method that can segment human behavior into each action and identify the action from the videos taken by a single camera. It uses features that can represent the shape of the human area in the depth direction, as well as the features such as motion direction and speed. The performance and effectiveness of the method are verified by experiments. Keywords: Behavior segmentation, Motion recognition, Optical flows, TMRIs, Ex-HOOF, MHI

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2024年02月22日  -  2024年02月25日

The most critical human sensory function resides in vision. This paper focuses on utilizing visual information, specifically self-perspective footage, to identify individual movements. Existing researches require third-party filming to recognize human body movements and states. The proposed method, on the other hand, simply attaches a camera to the human head and enables the recognition of the subject's actions. Consequently, it becomes easier to monitor daily movements of a human and gather his/her data on body kinetics. This approach would be beneficial in scenarios involving individuals engaging in risky behavior or, during a certain emergency, providing valuable assistance. Keywords: My VISION, Posture estimation, Optical flow, HSV conversion

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2024年02月22日  -  2024年02月25日

When individuals with visual impairment go out, public transportation such as trains and buses is commonly used. However, many of them experience accidents, such as falling from train platforms or tripping due to unexpected contact with other passengers. To solve this problem, we propose a method using the MY VISION system which detects obstacles that may pose risks to individuals with visual impairment. The proposed method detects obstacles such as passengers’ pillars and platform edges at train stations. We employ an RGB-D camera for capturing frontal views of a user, use depth images to detect the edge of obstacles and level differences, and give warning to the visually impaired user based on the distance between him/her and the detected obstacle. Experimental results show satisfactory performance of the method.

Keywords: Visually impaired, Safety, Train station platform, Obstacles detection, MY VISION

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2023年02月09日  -  2023年02月12日

Registration and segmentation of 3D data are necessary in many fields, such as factory automation, automated driving, or even in the medical field. However, the technique is generally applied to non-metal objects. One of the problems of registration of a metal object is that the point clouds representing a metal object contain many outliers and missing points because of its reflective nature. This makes the accuracy of the registration and segmentation degrade. In this paper, we propose registration and segmentation techniques that are robust to outliers. For registration, we use the globally optimal Go-ICP (Global optimal - Iterative Closest Point) algorithm considering the goodness of a combination of point cloud sets to escape from convergence to a local solution. In segmentation, we address the problem of RANSAC generating false segments consisting of nearly identical multiple planar points and introduce an improved RANSAC. We use three kinds of the metal tray to show the effectiveness of the proposed technique.

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2023年02月09日  -  2023年02月12日

In recent years, the development of automatic search methods based on aerial images taken by drones has been attracting attention in order to prevent secondary disasters and to perform rescue operations quickly in the search for victims of natural disasters. Although various methods exist for automatic person detection for search, they are based on the assumption that the background area of a person captured by a drone camera is a uniform ground in which only those persons who require rescue exist without any shadows or trees. In this paper, we propose a method of automatic detection of both fallen persons and person shadows, or trees on the ground. The method is combination of Ri-HOG and Ri-LBP features to search for fallen persons. These features are robust to rotation. We then employ GrabCut and brightness values to detect shadows. The effectiveness of the proposed method was verified by experiment.

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2023年02月09日  -  2023年02月12日

In recent years, the number of visually impaired persons has been increasing year by year, and outdoor accidents have also been increasing when they go out. It is difficult to detect hazards on sidewalks even with a currently popular technique, such as a semantic segmentation technique or YOLO, because sidewalk situations are complicated and change frequently. For this reason, we propose a method of recognizing sidewalk situations from a self-viewpoint video called MY VISION. Conventional methods detect objects surrounding the sidewalk by learning the objects’ features beforehand and guiding visually impaired persons according to the position/direction of the detected object. The proposed method neither learns objects nor detects objects. We focus on sidewalk situations and use a multi-class classification technique based on transfer learning of VGG-16 to guide visually impaired persons’ walk according to three kinds of sidewalk information to ensure more safety. The effectiveness of the proposed method was confirmed by experiments.

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Beijing   2022年11月17日  -  2022年11月19日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Online   2022年01月21日  -  2022年01月23日

For a visually impaired person, platforms of railway station are places where there is a risk of falling. This paper proposes a system that prevents fall of a visually impaired person from the platform by use of self-viewpoint images provided from MY VISION, an ego camera system attached to a user. The edges of the platform are detected by region segmentation and line segment detection on the self-viewpoint images. The performance of the proposed system was experimentally examined and satisfactory results were obtained.

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

In this paper, we propose a method of recognizing multiple objects using MSC-HOG (Multiple-Scale-Cell Histograms of Oriented Gradients) features and intensity models of both pedestrians and bicyclists. We also propose a method of detecting approaching passersby using different discriminators without using time-series information such as Optical Flow. The effectiveness of the proposed method is verified by experiments.

担当区分：筆頭著者,　責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper describes a method of navigating a visually impaired person who walks on a sidewalk and goes through a pedestrian crossing by the use of an ego-camera system named MY VISION. The proposed method finds a walkable region on the sidewalk. Thus, at a pedestrian crossing, it finds a traffic light and judges an appropriate timing to cross the road by analyzing the color and flashing status of the signal. The walkable region is found employing Graph- Based Segmentation (GBS), and once a pedestrian crossing is reached through the use of a GPS signal, a traffic light is detected by use of the HOG feature and Random Forest. The effectiveness of the proposed method was verified by experiments. This implies usefulness of a MY VISION system

DOI： 10.57417/jrnal.9.1_25

Scopus

CiNii Research

その他リンク： https://kyutech.repo.nii.ac.jp/records/7705

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tsukuba   2012年11月11日  -  2012年11月15日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Daejeon   2012年11月05日  -  2012年11月09日

DOI： 10.1007/978-3-642-37484-5_36

Scopus

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.1007/s00138-010-0298-4

Scopus

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In recent years, the number of visually impaired persons has been increasing year by year, and outdoor accidents have also been increasing when they go out. It is difficult to detect hazards on sidewalks even with a currently popular technique, such as a semantic segmentation technique or YOLO, because sidewalk situations are complicated and change frequently. For this reason, we propose a method of recognizing sidewalk situations from a self-viewpoint video called MY VISION. Conventional methods detect objects surrounding the sidewalk by learning the objects’ features beforehand and guiding visually impaired persons according to the position/direction of the detected object. The proposed method neither learns objects nor detects objects. We focus on sidewalk situations and use a multi-class classification technique based on transfer learning of VGG-16 to guide visually impaired persons’ walk according to three kinds of sidewalk information to ensure more safety. The effectiveness of the proposed method was confirmed by experiments.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85149473680&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In recent years, the development of automatic search methods based on aerial images taken by drones has been attracting attention in order to prevent secondary disasters and to perform rescue operations quickly in the search for victims of natural disasters. Although various methods exist for automatic person detection for search, they are based on the assumption that the background area of a person captured by a drone camera is a uniform ground in which only those persons who require rescue exist without any shadows or trees. In this paper, we propose a method of automatic detection of both fallen persons and person shadows, or trees on the ground. The method is combination of Ri-HOG and Ri-LBP features to search for fallen persons. These features are robust to rotation. We then employ GrabCut and brightness values to detect shadows. The effectiveness of the proposed method was verified by experiment.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85149470347&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Registration and segmentation of 3D data are necessary in many fields, such as factory automation, automated driving, or even in the medical field. However, the technique is generally applied to non-metal objects. One of the problems of registration of a metal object is that the point clouds representing a metal object contain many outliers and missing points because of its reflective nature. This makes the accuracy of the registration and segmentation degrade. In this paper, we propose registration and segmentation techniques that are robust to outliers. For registration, we use the globally optimal Go-ICP (Global optimal-Iterative Closest Point) algorithm considering the goodness of a combination of point cloud sets to escape from convergence to a local solution. In segmentation, we address the problem of RANSAC generating false segments consisting of nearly identical multiple planar points and introduce an improved RANSAC. We use three kinds of the metal tray to show the effectiveness of the proposed technique.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85149454546&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In recent years, the computer vision technology has been attracting more attention than ever and being applied in a wide range of fields. Among them, the technology on automatic recognition of human motion is particularly important, since it leads to automatic detection of suspicious persons and automatic monitoring of elderly people. Therefore, the research on human motion recognition using computer vision techniques has been actively conducted in Japan and overseas. However, most of the conventional researches on human motion recognition employs a video of a human motion taken using an external fixed camera. There is no research on human motion recognition using a video of a surrounding scenery provided from a wearable camera. This paper proposes a method of recognizing a human motion by estimating the posture change of a wearable camera attached to a walking human from the motion of a scenery in the video provided from the wearable camera and by analyzing a human trunk change obtained from the posture change of the camera. In the method, AKAZE is applied to the images to detect feature points and to find their correspondence. The 5-point algorithm is used to estimate the Epipolar geometry constraint and an essential matrix which provides a camera relative motion. The change of the camera relative motion is then used to analyze the shape of a human trunk. The analyzed results, i.e., walking motion features, are finally fed into a SVM to identify the motion. In the experiment, five types of walking motions are captured by a wearable camera from five subjects. The accuracy on human motion recognition was 80%. More precise feature points extraction, more exact estimation of motions, and considering variety of human walking motions are needed to improve the proposed technique.

DOI： 10.1145/3581807.3581809

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85160954407&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan  

As the global aging intensifies, it is more convenient for a robot to go for buying things like fruits and vegetables instead of elderly, and it is more human-like to select items according to a user’s personal preferences such as maturity of fruits, sweetness, etc. However, Fruits and vegetables are generally displayed in a disorderly manner. Therefore, detection and recognition of fruits and vegetables is a difficult task for a robot. This paper proposes an improved YOLOv3 and also pre-training the networks to detect fruits and vegetables, we then using Bilinear-CNN to classify fruit’s maturity. The effectiveness of the proposed method is shown by experiments.

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Online   2022年01月21日  -  2022年01月23日

In Japan, the population of the working-age between 15 and 64 years old is peaked in 1995 at about 87 million and is expected to continue to decline in the future. Therefore, in order to solve the labor shortage, the introduction of industrial robots that can perform the same level of work as humans is strongly requested.especially in a food preparation industry. In order to prevent danger to workers there, it is necessary for industrial robots to recognize workers and avoid them when there is fear of collision. In this paper, we propose a method of extracting hand regions based on the color distributions of a hand and GrabCut in an experimental environment to recognize human hands and detect their directions of approach. The performance of the proposed method is shown experimentally.

Kyutacar

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Online   2022年01月21日  -  2022年01月23日

In this paper, we propose a human-robot cooperative system to support shopping refugees. In the system, a robot acquires an object specified by a person in a distant site. The normal vector is calculated from the depth image, and the region is segmented using GBS on an RGB image. The two obtained clues are used to accurately detect the position of the specified object. The effectiveness of the proposed method was verified by experiments.

DOI： 10.5954/ICAROB.2022.OS21-1

Kyutacar

担当区分：筆頭著者,　責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

In Japan, the population of working age between 15 and 64 years old peaked in 1995 at about 87 million and is expected to continue to decline in the future. Therefore, to solve the labor shortage, the introduction of industrial robots that can perform the same level of work as humans is strongly requested especially in the food preparation industry. In order to prevent danger to workers there, industrial robots must recognize workers and avoid them when there is fear of collision. In this paper, we propose a method for extracting hand regions based on the color distribution of a hand and GrabCut in an experimental environment to recognize human hands and detect their directions of approach. The proposed method was examined experimentally and gave satisfactory results.

DOI： 10.57417/jrnal.9.3_289

Scopus

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85176549562&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

For a visually impaired person, platforms of railway station are places where there is a risk of falling. This paper proposes a system that prevents fall of a visually impaired person from the platform by use of self-viewpoint images provided from MY VISION, an ego camera system attached to a user. The edges of the platform are detected by region segmentation and line segment detection on the self-viewpoint images. The performance of the proposed system was experimentally examined and satisfactory results were obtained.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85125147415&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

As the global aging intensifies, it is more convenient for a robot to go for buying things like fruits and vegetables instead of elderly, and it is more human-like to select items according to a user’s personal preferences such as.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85125135095&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In this paper, we propose a method of recognizing multiple objects using MSC-HOG (Multiple-Scale-Cell Histograms of Oriented Gradients) features and intensity models of both pedestrians and bicyclists. We also propose a method of detecting approaching passersby using different discriminators without using time-series information such as Optical Flow. The effectiveness of the proposed method is verified by experiments.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85125137759&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In Japan, the population of the working-age between 15 and 64 years old is peaked in 1995 at about 87 million and is expected to continue to decline in the future. Therefore, in order to solve the labor shortage, the introduction of industrial robots that can perform the same level of work as humans is strongly requested.especially in a food preparation industry. In order to prevent danger to workers there, it is necessary for industrial robots to recognize workers and avoid them when there is fear of collision. In this paper, we propose a method of extracting hand regions based on the color distributions of a hand and GrabCut in an experimental environment to recognize human hands and detect their directions of approach. The performance of the proposed method is shown experimentally.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85125137027&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Hiroshima(Online)   2021年01月22日  -  2021年01月23日

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Hiroshima(Online)   2021年01月22日  -  2021年01月23日

Kyutacar

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Hiroshima(Online)   2021年01月22日  -  2021年01月23日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Hiroshima(Online)   2021年01月22日  -  2021年01月23日

Kyutacar

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108852379&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In this paper, we propose a system for a visually impaired person to cross a pedestrian crossing safely by the employment of the state of a traffic light and the remaining distance to the other side obtained from the analysis of MY VISION images. The traffic light at a pedestrian crossing is detected by a discriminator using HOG (Histograms of Oriented Gradients) features and Random Forest. The effectiveness of the proposed method was verified by experiments.

DOI： 10.5954/icarob.2021.os14-2

CiNii Research

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In order to search for victims in the event of a disaster, we propose a method of detecting a fallen person from UAV images and estimating its head position. Rotation-invariant HOG and Rotation-invariant LBP are used, so that they may detected a fallen person regardless of its body orientation. In addition, the position of the head of a fallen person is estimated using the peak of the gradient histogram. Experimental results show satisfactory performance of the proposed method.

DOI： 10.5954/icarob.2021.os14-1

CiNii Research

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In response to the requests from visually impaired people that they wish better assistance tools in their daily activities, and to the facts that they find difficulty in using public transportation, we propose a bus boarding support system using MY VISION. For the support, we detect a bus approaching a bus stop from MY VISION images employing Haar-like filters, and acquire the moving object area with the pyramidal Lucas-Kanade method. Then, we extract the frontal upper area displaying bus line number and the destination by random forest. Having obtained the area, a method of connecting area marking is used to obtain the bus line number area. The effectiveness of the proposed method is shown by experiments.

DOI： 10.5954/icarob.2021.os14-5

CiNii Research

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

One of the important roles of a camera surveillance system is to detect abnormal human actions or events. In this study, we propose a method of abnormal human actions/events detection method using Generative Adversarial Nets (GAN). In anomaly action detection, the main problem is that the image data of abnormal human actions is more difficult to obtain than normal human actions. To solve this difficulty, we use only normal human action data in the employed training network and those actions not recognized as normal are judged as abnormal. Experimental results show effectiveness of the proposed method.

DOI： 10.5954/icarob.2021.os14-3

CiNii Research

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

One of the important roles of a camera surveillance system is to detect abnormal human actions or events. In this study, we propose a method of abnormal human actions/events detection method using Generative Adversarial Nets (GAN). In anomaly action detection, the main problem is that the image data of abnormal human actions is more difficult to obtain than normal human actions. To solve this difficulty, we use only normal human action data in the employed training network and those actions not recognized as normal are judged as abnormal. Experimental results show effectiveness of the proposed method.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108852379&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In order to search for victims in the event of a disaster, we propose a method of detecting a fallen person from UAV images and estimating its head position. Rotation-invariant HOG and Rotation-invariant LBP are used, so that they may detected a fallen person regardless of its body orientation. In addition, the position of the head of a fallen person is estimated using the peak of the gradient histogram. Experimental results show satisfactory performance of the proposed method.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108801829&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In response to the requests from visually impaired people that they wish better assistance tools in their daily activities, and to the facts that they find difficulty in using public transportation, we propose a bus boarding support system using MY VISION. For the support, we detect a bus approaching a bus stop from MY VISION images employing Haar-like filters, and acquire the moving object area with the pyramidal Lucas-Kanade method. Then, we extract the frontal upper area displaying bus line number and the destination by random forest. Having obtained the area, a method of connecting area marking is used to obtain the bus line number area. The effectiveness of the proposed method is shown by experiments.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108842591&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In this paper, we propose a system for a visually impaired person to cross a pedestrian crossing safely by the employment of the state of a traffic light and the remaining distance to the other side obtained from the analysis of MY VISION images. The traffic light at a pedestrian crossing is detected by a discriminator using HOG (Histograms of Oriented Gradients) features and Random Forest. The effectiveness of the proposed method was verified by experiments.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108797473&origin=inward

担当区分：最終著者   記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）

<p><i>In recent years, aerial photography has been used to search for victims in the event of a disaster. Searching from the sky enables quick search activities in places that are difficult to enter. In this paper we propose a method of detecting a person fallen on the ground from images taken by a camera mounted on a UAV(multicopter). Unlike pedestrians, a fallen person takes various postures, and the orientation of the head in an image is not identical. Therefore, it is necessary to develop a method which is robust to various orientations of a fallen person. In the proposed method, Ri-HOG features and Ri-LBP features invariant to object orientation are employed for representing a fallen person, and the fallen person is detected by a classifier constructed using Random Forest. The effectiveness of the proposed method was verified by experiments</i><i>. </i></p>

DOI： 10.24466/pacbfsa.33.0_91

CiNii Article

CiNii Research

記述言語：英語   掲載種別：研究論文（学術雑誌）

Circulating tumor cells (CTCs) is a clinical biomarker for cancer metastasis. CTCs are cells circulating in the body of patients by being separated from primary cancer and entering into blood vessel. CTCs spread every positions in the body, and this is one of the cause of cancer metastasis. To analyze them, pathologists get information about metastasis without invasive test. CTCs test is conducted by analyzing the blood sample from patient. The fluorescence microscope generates a large number of images per each sample, and images contain a lot of cells. There are only a few CTCs in images and cells often have blurry boundaries. So CTCs identification is not an easy work for pathologists. In this paper, we develop an automatic CTCs identification method in fluorescence microscopy images. This proposed method has three section. In the first approach, we conduct the cell segmentation in images by using filtering methods. Next, we compute feature values from each CTC candidate region. Finally, we identify CTCs using artificial neural network algorithm. We apply the proposed method to 5895 microscopy images (7 samplesas), and evaluate the effectiveness of our proposed method by using leave-one-out cross validation. We achieve the result of performance tests, a true positive rate is 92.57% and false positive rate is 9.156%.

DOI： 10.1007/s11036-018-1121-0

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85053481889&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In recent years, the increase of aging population in developed countries has become a significant social problem. There are not a few accidents caused by their low walking ability. The improvement of the walking ability for the elderly and post recovery patients is the main concern of this paper. For this purpose, it is necessary to analyze and estimate their walking posture. In this paper, we propose a posture analysis and estimation method and a gait index acquisition method using a monocular wearable camera. The idea of this paper is to employ an ego-camera mounted on a user. His/her walking posture is analyzed by the frontal view video provided from the camera. This idea intends that the proposed system will be used on a daily basis in the future by those who want to improve their walking posture to keep a healthy life.

DOI： 10.1109/ICCCS49078.2020.9118537

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85087496874&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2020年01月13日  -  2020年01月16日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2020年01月13日  -  2020年01月16日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2020年01月13日  -  2020年01月16日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2020年01月13日  -  2020年01月16日

記述言語：英語   掲載種別：研究論文（学術雑誌）

In this article, a method to reduce artifacts on temporal difference images is introduced. The proposed method uses a nonrigid registration method for ground glass opacification (GGO), which is light in concentration and difficult to detect early. In this method, global matching, local matching, and three-dimensional (3D) elastic matching are performed on the current and previous images, and an initial temporal subtraction image is generated. After that, we use an Iris filter, which is the gradient vector concentration degree filter, to determine the initial GGO candidate regions and use supervoxel and graph cuts to segment region of interest in the 3D images. For each extracted region, a support vector machine is used to reduce the oversegmentation. The voxel matching is applied to generate the final temporal difference image, emphasizing the GGO regions while reducing the artifact.

DOI： 10.1109/MCE.2019.2941468

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85076374234&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

To realize a service-robot system for supporting a human life, we propose techniques for estimating self-position of a robot using local features of images, segmenting an image for finding a movable area, and also planning a route for finding the destination. In the route planning, junctions are labelled numbers for the robot to travel in order. The developed mobile robot travels to the destination employing the information on the estimated road region, its self-position and the planned route.

DOI： 10.5954/ICAROB.2020.GS5-2

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108847330&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a pedestrian detection method using a MY VISION system, which is an image processing system employing an ego-camera. First, a flow area different from the camera movement is extracted on the images from a user’s head-mounted camera. The area is a candidate area where pedestrians may exist. Second, MSC-HOG features are calculated to detect pedestrians. HOF feature is then computed to recognize the moving direction of the detected pedestrians. Experimental results show effectiveness of the proposed method.

DOI： 10.5954/ICAROB.2020.GS3-1

Scopus

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108810327&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

With the aggravation of the aging trend in Japan, the number of elderly people living alone has gradually increased, and the development of elderly people care system has begun to receive attention. Considering this trend, the paper proposes a MHI(Motion History Image)-based method called TMRIs (Triplet Motion Representation Images) that solve in recognition, a self-occlusive motion problem particularly in the depth direction using a single camera. The performance and effectiveness of the method are verified by experiments.

DOI： 10.5954/ICAROB.2020.GS3-2

Scopus

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108809693&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

A system is proposed that performs gait analysis of a pedestrian to prevent fall. In the system, a user walks with a chest-mounted camera. His/her walking posture is estimated using a pair of images obtained from the camera. Normally it is difficult to estimate the camera movement, when the parallax of the image pair is small. Therefore, the system uses a convolutional neural network. Optical flow and camera movement, and depth images are estimated alternately. Satisfactory results were obtained experimentally.

DOI： 10.5954/ICAROB.2020.GS1-2

Scopus

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85108785176&origin=inward

担当区分：最終著者,　責任著者   記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）

<p><i>In recent years, the labor shortage has become a serious problem in logistics industries. In particular, among the five major functions of logistics, most of the storage and cargo handling operations are performed manually using forklifts. This problem can be solved by automating the process. In this paper, we propose a method to acquire the position of cargo from the images taken by two fixed cameras in a warehouse and the relative distance information between a transport vehicle and the cargo using the camera attached to the vehicle. The effectiveness of the proposed methods is shown by experiments.</i><i> </i></p>

DOI： 10.24466/pacbfsa.32.0_b1-3

CiNii Article

CiNii Research

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Bangladesh   Rajshahi   2019年07月11日  -  2019年07月12日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Salient region of an image is usually detected by using contrast and boundary priors. Along with those cues the use of seam importance map has shown promising output previously. In this study, better result is found by further exploiting the seam-map using spatial distance and color information in combination with boundary prior. Color and seam maps are also down-weighted using average spatial distance to other regions. Moreover, passing the superpixelized version of the input image into seam and color map generation procedure has improved the output. Experimental results based on MSRA 1k dataset are presented with ten state of the art methods. F-beta measures are presented along with precision recall curves to better understand the outcome. The performance comparison with compared researches proofs superiority of the proposed method.

DOI： 10.1109/IC4ME247184.2019.9036637

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85082984951&origin=inward

担当区分：最終著者,　責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

As action recognition undergoes change as a field under influence of the recent deep learning trend, and while research in areas such as background subtraction, object segmentation and action classification is steadily progressing, experiments devoted to evaluate a combination of the aforementioned fields, be it from a speed or a performance perspective, are far and few between. In this paper, we propose a deep, unified framework targeted towards suspicious action recognition that takes advantage of recent discoveries, fully leverages the power of convolutional neural networks and strikes a balance between speed and accuracy not accounted for in most research. We carry out performance evaluation on the KTH dataset and attain a 95.4% accuracy in 200 ms computational time, which compares favorably to other state-of-the-art methods. We also apply our framework to a video surveillance dataset and obtain 91.9% accuracy for suspicious actions in 205 ms computational time.

DOI： 10.1007/s10015-018-0518-y

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85071613991&origin=inward

担当区分：筆頭著者,　責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper proposes a method of analyzing a human walk environment using MY VISION. MY VISION is an ego-camera and a computer system which analyzes a video obtained from the ego-camera to acquire certain visual information useful for hu- man daily activities. The system is expected to be a virtual eye of a visually impaired person or the third eye of a pedestrian absorbed in a mobile phone. The proposed method keeps in a database the background images of key points along a sidewalk and judges if a MY VISION user is walking along the sidewalk or if he/she has come to a crosswalk by referring to the backgrounds. If the former, the method finds a safe road region on the sidewalk, whereas, if the latter, it searches for the crosswalk for finding an appropriate walk direction and a traffic light to know the proper timing to cross it. Experimental re- sults show effectiveness of the proposed method.

DOI： 10.24507/ijicic.15.03.861

Scopus

CiNii Article

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85067556972&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Most of the saliency detection methods use the contrast and boundary prior to extract the salient region of an input image. These two approaches are followed in Boundary Aware Regional Contrast Based Visual Saliency Detection (BARC) [1] along with spatial distance information to achieve state of the art result. In this research, a more interesting cue is introduced to extract the salient region from an input image. Here, a combination of seam map and BARC [1] is presented to produce the saliency output. Seam importance map with boundary prior is also presented to measure the performance of this combination. Experiments with ten state of the art methods reveal that we get better saliency output by combining seam information of an input image with BARC [1].

DOI： 10.1109/CIET.2018.8660825

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85063582435&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Singapore   Singapore   2019年01月06日  -  2019年01月09日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Singapore   Singapore   2019年01月06日  -  2019年01月09日

Kyutacar

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（研究会，シンポジウム資料等）

<p><i>A method of 3-D shape recovery employing uncalibrated cameras is described. Three-dimensional shape recovery has long been studied to date and has various application fields. An established method for 3-D shape recovery is based on stereo cameras which necessitates camera calibration in advance. In this paper, a 3-D shape recovery method by uncalibrated cameras is presented, which ranges from the recovery of partial shape of an object to its entire shape. One of the main advantages of the present method is that the shape recovery is not restricted to a video capture site such as a studio, if only videos are able to be taken by cameras.</i><i> </i></p>

DOI： 10.24466/pacbfsa.32.0_B1-1

CiNii Article

CiNii Research

その他リンク： https://ci.nii.ac.jp/naid/130007980073

担当区分：筆頭著者,　責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

<p><i>This paper proposes a new method of describing a self-occlusive human motion, particularly in the depth direction, which has been considered little in the motion/action recognition studies to date in spite of its importance in our daily life. A Motion History Image (MHI) is a well-known method of describing a motion by a single gray value image, but it suffers from a self-occlusion problem in which present motion overwrites past motion. To solve this difficulty, a Reverse description MHI</i>(RMHI) is proposed in the paper. RMHI and the original MHI are both employed for motion representation in the proposed method; the former for approach motion, whereas the latter for leave motion. In the experiment on motion recognition, motions are described by RMHI or MHI according to motion direction, transformed then to Hu moment vectors, and finally recognized employing the k-nearest neighbor. Experimental results show effectiveness of the RMHI description.</p>

DOI： 10.24466/ijbschs.24.1_1

CiNii Article

CiNii Research

その他リンク： https://ci.nii.ac.jp/naid/130007998971

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In recent years, along with the aging of society worldwide, decrease of working population has become a serious problem. For this reason, robots are expected to substitute human work, automate distribution, and support human daily life, especially, elderly care assistance. In this research, we focus on the support of those who need care in everyday life, and, in this paper, we propose a human-robot cooperative system that supports acquisition of objects in cooperation with a human. The outline of the action of this robot is as follows: (i) It moves to the location designated by a user autonomously (not by remote control): (ii) On arrival, the robot exchanges information by video with the user who is at a remote place, and acquires the objects designated by the user among those placed there. (iii) After the acquisition, the robot moves again autonomously to the user and hands over the objects to the user. In this paper, we focus on step (ii) and show methods and some experimental results.

DOI： 10.1117/12.2522001

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85063913144&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Bangladesh   Dhaka   2018年12月27日  -  2018年12月29日

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In Japan, the number of patients with osteoporosis and rheumatoid arthritis is increasing. Image diagnosis using CR images is effective for osteoporosis and rheumatoid arthritis. Development of a CAD system is important for reducing burdens on doctors. In this paper, we propose an automatic registration algorithm in the CAD system. In the proposed method, the genetic algorithm is used to register bone regions between identical parts of the same subject with different time series. In the experiment, the proposed method is applied to 176 bone area, and 98.14 % of TPR, 1.85 % of FPR are obtained respectively. Even when the area difference is used as the fitness of the genetic algorithm, it has cross-correlation and positioning accuracy equivalent to mutual information.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85060451078&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Osteoporosis is one of the famous bone diseases. It is a major cause of deteriorating the quality of life, and early detection and early treatment are becoming socially important. Visual screening using Computed Radiography (CR) images is effective for diagnosis of osteoporosis, but there are problems of increasing the burden on doctors, variation in diagnostic results due to differences in experiences of doctors, and undetected lesions. In order to solve this problem, we are working on a computer-aided diagnosis (CAD) system for osteoporosis. In this paper, we propose segmentation methods of the phalange region from the phalangeal CR images as a preprocessing of classification of osteoporosis. In the proposed method, we construct a segmentation model using U-Net, which is a type of deep convolution neural network (DCNN). The proposed method was applied to input images generated from CR images of 101 patients with both hands, and evaluated using the Intersection over Union (IoU) values. The result was 0.914 in IoU.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85060468143&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In recent years, dysphagia is problem among elderly people. Therefore, it is necessary to accurately evaluate swallowing function in order to prevent swallowing disorder beforehand or to detect it early. And it is considered that evaluation of swallowing function using Magnetic Resonance Imaging (MRI) is useful. In order to accurately analyzing of the swallowing motion using a computer aided diagnosis (CAD) system on MR imaging, automatic extraction of the esophagus region, which is a region of interest by the image analysis method, is required. Extraction of the spinal region is required as a preliminary step of the esophagus region extraction. Therefore, in this paper, we develop an analysis method of swallowing movement by three steps of extraction of spinal region, extraction of esophageal region, and analysis of swallowing movement. As an analytical method of swallowing movement, we emphasize the liquid part at the time of swallowing movement using an emphasis map, then follow the liquid tracing by using the AKAZE feature quantity and the particle filter algorithm, and analyze the swallowing motion.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85060441189&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Cleft lip is a kind of congenital facial morphological abnormality. In the clinical field of cleft lip, it is necessary to analyze symmetric shape. However, there is no method to analyze the cleft lip technique based on symmetrical viewpoints. On the other hand, in our previous method to find a symmetric axis using a 2D image, since the middle line is extracted only from the front view of the face moire image. There was a problem that low accuracy was obtained by slight rotation of the face and it was not possible to consider 3D information. In this paper, we propose a method to extract the median plane of the face by analyzing based on bilateral symmetry by using 3D point cloud on the face of front. By extracting the median plane, we believe that not only surgical assistance of doctor be possible but also become a clue to development of simulation software which is the end goal.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85060498535&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   PyeongChang   2018年10月17日  -  2018年10月20日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   PyeongChang   2018年10月17日  -  2018年10月20日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   PyeongChang   2018年10月17日  -  2018年10月20日

記述言語：英語   掲載種別：研究論文（学術雑誌）

In this paper, we propose a method to reduce artifacts on temporal difference images by improving the conventional method using a non-rigid registration method for ground glass opacification (GGO), which is light in concentration and difficult to detect early. In this method, global matching, local matching, and 3D elastic matching are performed on the current image and past image, and an initial temporal difference image is generated. After that, we use an Iris filter, which is the gradient vector concentration degree filter, to determine the initial GGO candidate regions and perform segmentation using SuperVoxel and Graph Cuts in which a superpixel is extended to three dimensions for each region of interest. For each extracted region, a support vector machine (SVM) is used to reduce the over-segmentation. Finally, in the method that greatly reduces artifacts other than the remaining GGO candidate regions, Voxel Matching is applied to generate the final temporal difference image, emphasizing the GGO regions while reducing the artifact. The resulting ratio of artifacts to lung volume is 0.101 with an FWHM of 28.3, which is an improvement over the conventional method and shows the proposed method’s effectiveness.

DOI： 10.1007/s11036-018-1111-2

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85053445997&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Tumor of spinal cord (Spinal Cord Neoplasms (SCN)) is considered as one of the life threatening diseases that causes death. Early detection of the SCN plays an important role in the management of the lesions. To analyze the treatment, it is necessary to segment the spinal canal based on accurate three-dimensional image processing technique. This paper presents a segmentation algorithm based on 3D region growing for extracting spinal canal from CT images with high accuracy. Intersection over union (IoU) is used to compare the results of segmentation with the manual segmentation results. In the experiment, the proposed method was tested on 3373 CT slices of 10 patients. The proposed method has an average accuracy of 0.7732 and a variance of 0.0061. Satisfactory results have been achieved rapidly, which demonstrates the effectiveness and superiority of the proposed method.

DOI： 10.1109/ICT-ROBOT.2018.8549913

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85060055563&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Since the declining birthrate and the aging of society, there is concern about the labor shortage in Japan. There is a movement to compensate for the labor shortage by automation of factories by robots. Automation technique is wildly promoted in logistics industry, while there is few studies in objects picking. To solve this issue, we develop an image detection scheme for robotics picking from a video image. It is difficult to recognize and grasp different types of objects in robot vision field. Therefore, in the proposed method, object detection and object recognition method are proposed using Region-based Fully Convolutional Networks that is a type of object detection using deep learning. After detecting the object individually, final target object can select by applying the GrowCut algorithm. As a result, we achieve 0.6773 of the average precision and 0.6395 of Intersection over Union as the segmentation result respecively.

DOI： 10.1109/ICT-ROBOT.2018.8549879

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85060010049&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Busan   2018年09月06日  -  2018年09月08日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Busan   2018年09月06日  -  2018年09月08日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In this paper, we propose a FTOP (Feature Transform Optimization Problem) and its solution. We propose a method to optimize both parameters and processing order of feature transform simultaneously, not limited to convolution and pooling included in CNN (Convolutional Neural Network). In order to realize the optimization, we formulate it as a combinatorial optimization problem and solve it by meta-heuristics. The effectiveness of the proposed method is shown by applying the proposed method to pedestrian classification based on a benchmark data set.

DOI： 10.23919/SICE.2018.8492633

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85056737437&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   PyeongChang   2018年10月17日  -  2018年10月20日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

日本   奈良   2018年09月11日  -  2018年09月14日

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2018, © 2018 Informa UK Limited, trading as Taylor & Francis Group. This paper presents a method to recognise actions, which are overlapping and multi-dimensionalities. A spatio-temporal representation is illustrated on local interest points to compute global features. Motion history image (MHI) is computed and motion overwriting the motion overwriting problem of the MHI. The main contribution of this paper is that it demonstrates a higher discriminative ability of various complex actions when compared to the other MHI-based approaches. It selects local interest feature points to capture motion information using Speeded-Up Robust Features (SURF). These key interest points are exploited to compute gradient-based optical flow into four channels. RANSAC is exploited to remove outliers. It incorporates frame-subtracted accumulated image so that we can mask out points that are not required. Afterwards, feature vectors are computed based on moments. Actions are recognised by employing a nearest neighbour classifier and leave-one-out cross-validation partitioning scheme. The proposed method provides satisfactory recognition rates over several other approaches for some challenging actions in outdoor scenes.

DOI： 10.1080/21681163.2017.1298472

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85016247850&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

At present, the aging population is growing in Japan. Along with that, the need for the utilization of welfare equipment is increasing. Electric wheelchair, a convenient transportation tool, is popularized rapidly. However, many accidents have occurred by using electric wheelchair, and the dangers for driving are pointed out. Therefore, it needs to improve accident factors, reduce accidents and improve the convenience of electric wheelchair by automation. Environmental recognition is the key technology for developing autonomous electric wheelchair. Environmental recognition includes self-position estimation, recognition of sidewalks, crosswalks, traffic lights, and moving object prediction, etc. In order to solve these problems, this paper describes a system for the detection of sidewalks, crosswalks and traffic lights. We develop the object recognition methods using a modified YOLOv2 that is one of object detection algorithms applying convolutional neural networks (CNN). We detect the object through YOLOv2 and perform processing such as unnecessary bounding box deletion and interpolation. The experimental results demonstrate that the area under the curve (AUC) of the detection rate is 0.620.

DOI： 10.1145/3278229.3278231

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058546828&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2018年08月22日  -  2018年08月24日

担当区分：責任著者   記述言語：日本語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（学術雑誌）

Although radiographic assessment of joint damage is essential in characterizing disease progression and prognosis in patients with rheumatoid arthritis (RA), it is often difficult even for trained radiologists to find radiographic changes on hand and foot radiographs because lesion changes are often subtle. This paper proposes a novel quantitative method for automatically detecting bone erosion on hand radiographs to assist radiologists. First, the proposed method performs with the crude segmentation of phalanges regions from hand radiograph and extracts the detailed phalanges regions by the multiscale gradient vector flow (MSGVF) Snakes method. Subsequently, the region of interest (ROI; 40 × 40 pixels) is automatically set on the contour line of the segmented phalanges by the MSVGF algorithm. Finally, these selected ROIs are identified by the presence or absence of bone erosion using a deep convolutional neural network classifier. This proposed method is applied to the hand radiographs of 30 cases with RA. The true-positive rate and the false-positive rate of the proposed method are 80.5% and 0.84%, respectively. The number of false-positive ROIs is 3.3 per case. We believe that the proposed method is useful for supporting radiologists in imaging diagnosis of RA.

DOI： 10.1007/s11042-017-5449-4

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85056120147&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 IEEE. In this paper, a method is proposed for detecting and tracking a specific moving object (e.g., a bus) on the road from images of a camera attached to the head of a user, aiming at developing a system to support daily lives of visually impaired people. The proposed method traces feature points on the images, extracts a moving object region, and detects a bus by applying Haar-like feature and random trees to the region. The effectiveness of the proposed method is shown experimentally.

DOI： 10.1109/SII.2017.8279323

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85050871522&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

日本   大分   2018年01月18日  -  2018年01月20日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

日本   大分   2018年01月18日  -  2018年01月20日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

日本   大分   2018年01月18日  -  2018年01月20日

記述言語：英語   掲載種別：研究論文（学術雑誌）

骨のおもな疾患として，骨粗しょう症が挙げられる．骨粗しょう症に対する画像診断は有効であるが，類似した低骨量を呈する画像も多く，画像診断における客観性や再現性の問題がある．そこで本稿では，指骨computed radiography（CR）画像から骨粗しょう症の自動識別手法を提案する．提案手法では，深層畳み込みニューラルネットワーク（DCNN）を用いた識別器を構築し，骨粗しょう症有無の識別を行う．DCNNの学習および識別には，CR画像から3種類の画像を作成し，各指骨領域内部からROIを抽出後，この3種類のROIをR，G，Bチャンネルに割り当て生成した疑似カラー画像を用いる．実験では，101症例に対し提案手法を適用し，真陽性率（TPR）：75.5［％］，偽陽性率（FPR）：13.9［％］という結果を得た．

DOI： 10.11409/mit.36.90

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006588793

担当区分：筆頭著者   記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）

<p><i>This paper proposes a new method of describing a human motion based on Motion History Image (MHI) representation. MHI representation is a well-known way of describing a motion by a single image, but it cannot deal with a self-occlusive motion in a proper way. This paper focuses on the motion in depth direction and proposes Reverse Descriptive MHI (RD-MHI). The RD-MHI and the original MHI are both employed for motion recognition, since the former well describes the motion toward an observing camera, whereas the latter away from the camera. Experimental results support the combination in human motion recognition.</i><i><b> </b></i></p>

DOI： 10.24466/pacbfsa.31.0_25

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130007979978

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a system for caring a physically impaired person by a mobile robot employing cooperative object recognition. It is a well-known fact that object recognition is not an easy task for a robot particularly in a real environment. It is, on the contrary, easy for a human. In the proposed system, a mobile robot goes to a spot where a user wishes to go in place of him/her and transfers images containing various objects in the spot. Having seen the images, the user chooses one of those objects and sends a request to the robot to get it. The robot performs the request by taking and bringing it back to the user. As an example, the robot is asked to go to a store, buy goods which a user wants and bring them back to the user. In this process, the goods in the captured images are segmented by the robot and they are recognized and one of them is chosen by the user. In this way, cooperative object recognition is realized to achieve a certain purpose. A laboratory experiment was done with promising results. The idea is different from remote control of a mobile robot, because the robot is autonomous and a user takes a role of only recognizing objects and sending his/her request to the robot.

DOI： 10.1109/TENCON.2017.8228030

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044202825&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. Circulating tumor cells (CTCs) is a useful biomarker for cancer metastasis. The blood from a cancer patient is analyzed by a fluorescence microscope. Each case takes a large number of images, which usually have a lot of cell regions. Thus, analyzing the images is hard work for pathologists, and misdiagnosis may happen. In this paper, we develop an automatic CTCs identification method for fluorescence microscopy images. The proposed method consists of three steps. First, we extract cell regions in images using filtering methods. Second, we compute features of each CTC candidate regions. Finally, we identify the CTCs using AdaBoost algorithm. And we analyze the features to know which ones are effective for characterizing CTCs and normal cells. We apply the proposed method to 5040 microscopy images, and evaluate the effectiveness of our method by using leave-one-out cross validation. We achieve a true positive rate of 97.30 [%] and a false positive rate of 12.82 [%].

DOI： 10.23919/ICCAS.2017.8204219

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044436784&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. In recent years, the development of the computer-aided diagnosis (CAD) systems to support radiologist is attracting attention in medical research field. Temporal subtraction technique, which is one of CAD, is a technique to generate images emphasizing temporal changes in lesions by performing a differential operation between current and previous image of the same subject. In this paper, we propose an image registration method for image registration of current and previous image, to generate temporal subtraction images from CT images and enhanced bone metastasis region. The proposed registration method is composed into three main steps: i) segmentation of the region of interest (ROI) using graph cut, ii) use global image matching to select pairs from previous and current image, and iii) final image matching based on salient region feature. We perform our proposed method to synthesis and satisfactory registration experiments. The rotated synthesis image give TP 100.0[%] and FP 12.16[%]. The synthesis image obtained by applying a Gaussian filter give TP 70.40[%] and FP 0.00[%]. The synthesis image obtained by adding artificial pseudo lesion region give TP 99.45[%] and FP 17.89[%]. The synthesis image obtained by adding random noise of 5[%], which gave TP 83.05[%] and FP 16.95[%].

DOI： 10.23919/ICCAS.2017.8204242

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044465807&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. About 4% of babies in Japan suffer from congenital anomalies. Cleft lip is the most common disorder among external malformations occurring at a rate of 1 in 500 people. The goal of treatment is to recover symmetric and functional lips and nose forms. However, in the case of unilateral cleft lip where the lips and nose themselves are shifted from the midline of the face, it is difficult to set the midline which is the symmetry axis for evaluating the degree of symmetry of the face. In this paper, we propose an image processing method for extracting the midline using zebra image and shadow image which are two dimensional image created based on three dimensional analysis. Also, we evaluate the degree of asymmetry of the face by using the midline as the axis. As a result of applying the proposed method to 25 cases, which indicated that the proposed method is useful.

DOI： 10.23919/ICCAS.2017.8204239

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044463365&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. In recent years, dysphagia is problem among elderly people. Therefore, it is necessary to accurately evaluate swallowing function in order to prevent swallowing disorder beforehand or to detect it early, and it is considered that evaluation of swallowing function using MRI is useful, and the demand for development of CAD system using MRI is increasing. In order to accurately analyze the swallowing motion, automatic extraction of the region of interest by the image analysis method is necessary. In this paper, as a pretreatment of a method for automatic extraction of esophageal region on dental MR image, we develop automatic extraction method of spinal candidate region. We first extract the posterior region of the spine and then extract the intervertebral disc region. Then we perform the detection of the anterior region of the spine and finally we extract the region between the anterior region of the spine and the posterior region of the spine as a spinal region. The proposed method was applied to 5 cases of dental MR images, and the results of TP 91.2 [%] and FP 8.8 [%] are obtained.

DOI： 10.23919/ICCAS.2017.8204243

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044477204&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. Lung cancer is the leading cause of death which accounts for the number of deaths in cancer in the world. Early detection and early treatment are regarded as an important. Especially, the ground glass opacity (GGO) is a shadow called pre-cancerous lesion, but it is a shadow which is difficult to detect by a radiologist because of haze and complicated shape. Therefore, in recent years, a computer aided diagnosis (CAD) system has been developed for the purpose of improving the detection accuracy for early detection and reducing the burden to radiologists. In this paper, we extract the GGO using Deep Convolutional Neural Network (DCNN) based on emphasized images. Before detect a GGO region, we apply preprocessing such as isotropic voxel to the original images, and extraction of the lung area. Next, we remove the vessel and bronchial region by 3D line filter based on Hessian matrix, and extract the initial candidate regions using density gradient, volume and sphericity. Subsequently, we segment the candidate regions, extraction of features, and reducing false positive shadows. Finally we create emphasize images and identify with DCNN using those images. As a result of applying the proposed method to 31 cases on Lung Image Database Consortium (LIDC), we obtained a true positive rate (TP) of 86.05 [%] and false positive number (FP) of 4.81 [/case].

DOI： 10.23919/ICCAS.2017.8204464

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044442253&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. The Japanese population has rapidly aged and the number of aged persons who have lower physical ability has increased recently. Thus the development of medical and healthcare devices is expected. Wheelchair requires care support in most cases. Therefore the development of autonomous wheelchair is meaningful since we can expect to improve convenience and to reduce burden of caregivers. The autonomous wheelchair requires several techniques. Our research is to develop a navigation system based on image processing techniques. However, we assume that the system instructs an appropriate direction to head towards the destination when a wheelchair user comes to a crossing. Incidentally, deep learning, a kind of artificial neural network, has attracted attention in the field of machine learning in recent years. This paper proposes methodology for supporting autonomous driving by use of a classifier trained on a video images with deep learning. Also, we apply visual odometry to generate training data.

DOI： 10.23919/ICCAS.2017.8204217

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044473264&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. In recent years, the death rate caused by lung cancer is increasing. To detect the lung cancer, multi detector-row computed tomography (MDCT) images are used in visual screening. Lung cancer can be easily detected by using the chest MDCT images, however, it has enormous images and burden to radiologists. Research and development of the computer aided diagnosis (CAD) system have been assisted the diagnosis. As one of the CAD technologies, temporal subtraction technique is possible to emphasize the changing interval on the CT images. It uses subtraction operation between previous and current CT images of the same patient. On the other hand, pattern recognition using image reconstruction by sparse coding method has attracted attention. This technique is mathematically modeling the information processing by the primary visual cortex of human. It is the technique for representing images by the linear combination of a small number of basis. In this paper, candidate nodules under 20[mm] were segmented from temporal subtraction images based on the 3D sparse coding technique.3D sparse coding is three dimensional expansion of the sparse coding. Also, we classified the final candidate nodules using support vector machine (SVM) method based on coefficient matrix which are obtained by the 3D sparse coding. We applied proposed method to 31 cases of chest MDCT images in which the number of nodules was more than one. We achieved experimental result with true positive rates (TPR) of 70.2[[%], and false positive rates (FP) of 34.7[/scan], respectively.

DOI： 10.23919/ICCAS.2017.8204220

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044482344&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Control, Robotics and Systems - ICROS. Segmentation of liver from Multi-phase CT images is one of the essential technology for computer aided diagnosis. Contrast medium gives multi-phase CT images different intensity feature which enables to detect tumor. It is a challenging problem to segment liver region from multi-phase CT images. There are many approaches for solving this problem, however, these methods depend on other phases or registration. In order to solve this problem, we propose anatomy feature-based method which is mostly independent for each phase in this paper. This method uses level set method for final segmentation. The accuracy of segmentation result by level set methods relay on initial contour, so we preprocess initial region of liver by anatomical feature. Then we introduced contour constrain by using ribs information to improve segmentaion accuracy. Our segmentation was evaluated on 5 multi-phase CT images which have 4 phases. Experimental results show that the proposed method is good accuracy for each phase.

DOI： 10.23919/ICCAS.2017.8204240

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044443256&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Osteoporosis is known as a disease of bone. Visual screening using Computed Radiography (CR) images is an effective method for osteoporosis, however, there are many similar diseases that exhibit state of low bone mass. In this paper, we propose an automatic identification method of osteoporosis from phalanges CR images. In the proposed method, we implement a classifier based on Deep Convolutional Neural Network (DCNN), and identify unknown CR images as normal or abnormal. For training and evaluating of CNN, we use pseudo color images. In the experiment, we apply our proposal method to 101 cases and TPR of 64.7 [%] and FPR of 6.51 [%] were obtained.

DOI： 10.23919/ICCAS.2017.8204241

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044473373&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Cancer is a leading cause of death both in Japan and worldwide. Detection of cancer region in CT images is the most important task to early detection. Recently, visual screening based on CT images become useful tools for cancer detection. However, due to the large number of images and the complexity of the image processing algorithms, image processing technique is still required a high screening quality. To overcome this problem, some computer aided diagnosis (CAD) algorithms are proposed. In this paper, we have designed and developed a framework combining machine learning based on deep convolutional neural networks (DCNN) and temporal subtraction techniques based on non-rigid image registration algorithm. Our main classification method can be built into three main steps; i) pre-processing for image segmentation, ii) image matching for registration, and iii) classification of abnormal regions based on machine learning algorithms. We performed our proposed technique to 25 thoracic MDCT sets and obtained true positive rates of 92.31 [%], false positive rates of 6.32 [/case] were obtained.

DOI： 10.23919/ICCAS.2017.8204218

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044482182&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taipei   2017年12月11日  -  2017年12月14日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2017 The Society of Instrument and Control Engineers - SICE. According to the increase in the number of elderly people living alone, nursing care systems for such people have become much more important than ever. This paper focuses its attention on the detection of abnormal motions such as falling of the elderly by use of a camera settled in a room. Although there are several human motion representation and recognition methods, they all deal with the motions acted in a plane perpendicular to camera view, which suggests weakness of their motion description method. This paper proposes a novel method of describing a human motion independent of motion direction. The method expands the original 2-D Motion History Image to a 3-D version and computes the Hu moments in a 3-D way for the motion recognition. The method as well proposes a set of three 2-D images for motion description derived from the 3D-MHI to reduce computational load. The performance of the proposed method is shown experimentally and discussion is given.

DOI： 10.23919/SICE.2017.8105563

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85044144989&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Malaysia   Penang   2017年11月05日  -  2017年11月08日

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2017, CARS. Purpose: A temporal subtraction (TS) image is obtained by subtracting a previous image, which is warped to match the structures of the previous image and the related current image. The TS technique removes normal structures and enhances interval changes such as new lesions and substitutes in existing abnormalities from a medical image. However, many artifacts remaining on the TS image can be detected as false positives. Method: This paper presents a novel automatic segmentation of lung nodules using the Watershed method, multiscale gradient vector flow snakes and a detection method using the extracted features and classifiers for small lung nodules (20 mm or less). Result: Using the proposed method, we conduct an experiment on 30 thoracic multiple-detector computed tomography cases including 31 small lung nodules. Conclusion: The experimental results indicate the efficiency of our segmentation method.

DOI： 10.1007/s11548-017-1598-1

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85019139416&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Circulating tumor cells (CTCs) are a useful biomarker since they may have some information about cancer metastasis. The blood from cancer patient is analyzed by a fluorescence microscope. It takes a large number of photos for each case, and many cells are contained in the microscopy images. Thus, analyzing them is hard work for pathologists. This work tends to depend on the individual skill of pathologist so misdiagnosis may be happen. In this paper, we develop an automatic CTCs identification method in fluorescence microscopy images based on artificial neural network. We applied our proposed method to 5040 microscopy images (6 cases), and evaluated the effectiveness of our method by using leave-one-out cross validation. We achieve a true positive rate of 98.65 [%] and a false positive rate of 18.24 [%].

DOI： 10.1145/3133793.3133798

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85052677091&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

記述言語：日本語   掲載種別：研究論文（学術雑誌）

担当区分：筆頭著者,　責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2016 IEEE. This paper describes a portable system for finger-spelling recognition, employing a PC and a video taken by an ego-camera mounted on the body of a person who does finger-spelling. The system is intended to be a useful tool for an orally impaired person to communicate to anyone at any place by carrying it with him/her. The images of the finger-spelling hand of a user who is carrying the system is captured by an ego-camera. The hand is extracted from arbitrary backgrounds by use of a Gaussian mixture model and skin color evaluation: The trimmed and normalized image of the extracted hand is recognized employing the feature space defined by applying the principal component analysis to the learning data containing 45 finger-spelled Japanese Hiragana letters each with 50 samples and the nearest neighbor method. The on-line performance of the proposed system is experimentally shown.

DOI： 10.1109/TENCON.2016.7848021

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85015432805&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2016 Institute of Control, Robotics and Systems - ICROS. In recent years, development of the computer-aided diagnosis (CAD) systems for the purpose of reducing the false positive on visual screening and improving accuracy of lesion detection has been advanced. Lung cancer is the leading cause of cancer death in the world. Among them, GGO (Ground Glass Opacity) that exhibited early in the before cancer lesion and carcinoma in situ shows a pale concentration, have been concerned about the possibility of undetected on the screening. In this paper, we propose an automatic extraction method of GGO candidate regions from the chest CT image. Our proposed image processing algorithms is consist of four main steps; 1) segmentation of volume of interest from the chest CT image and removing the blood vessel regions, bronchus regions based on 3D line filter, 2) first detection of GGO regions based on density and gradient which is selected the initial GGO candidate regions, 3) identification of the final GGO candidate regions based on DCNN (Deep Convolutional Neural Network) algorithms. Finally, we calculates the statistical features for reducing the false-positive (FP) shadow by the rule-based method, performs identification of the final GGO candidate regions by SVM (Support Vector Machine). Our proposed method performed on to the 31 cases of the LIDC (Lung Image Database Consortium) database, and final identification performance of TP: 93.02[%] , FP: 128.52[/case] are obtained respectively.

DOI： 10.1109/ICCAS.2016.7832398

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85014043688&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2016 Institute of Control, Robotics and Systems - ICROS. Circulating tumor cells (CTCs) is an informative biomarker which assists pathologists in early diagnosis and evaluating therapeutic effects of patients with malignant tumors. The blood from a cancer patient is analyzed by a microscope and a large number of pictures including many cells are generated for each case. Thus, analyzing them is time-consuming work for pathologists, and misdiagnosis may happen since the diagnosis of CTCs tends to depend on the individual skill of pathologist. In this paper, we propose a method which detects cell candidate regions in microscopy images automatically to make quantitative analysis possible by computer. Our proposed method consists of three steps. In the first step, we extract initial cell candidate regions in microscopy images based on the saliency map. In the second step, we choose non-single cell regions from the initial candidates based on the SVM algorithm. In the third step, we separate connected regions into single cell regions based on the branch and bound algorithm. We demonstrated the effectiveness of our proposed method using 540 microscopy images and we achieved a true positive rate of 99.04[%] and a false positive rate of 3.95[%] .

DOI： 10.1109/ICCAS.2016.7832397

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85014007441&origin=inward

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

日本   大分   2017年01月19日  -  2017年01月21日

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Detection of objects from a video is one of the basic issues in computer vision study. It is obvious that moving objects detection is particularly important, since they are those to which one should pay attention in walking, running, or driving a car. This paper proposes a method of detecting moving objects from a video as foreground objects by inferring backgrounds frame by frame. The proposed method can cope with various changes of a scene including large dynamical change of a scene in a video taken by a stationary/moving camera. Experimental results show satisfactory performance of the proposed method.

DOI： 10.1007/s10015-016-0347-9

Kyutacar

Scopus

その他リンク： https://link.springer.com/article/10.1007/s10015-016-0347-9

記述言語：英語   掲載種別：研究論文（学術雑誌）

<p>Circulating tumor cells（CTCs）can be a useful biomarker. They may have some information about the malignant disease, since they are one of causes of the cancer metastasis. The blood sample from cancer patient is analyzed by fluorescence microscope. This microscope takes enlarged images with three types of lights（red, green and blue）,and specific materials are reacted respectively. The blood contains a lot of cells, but there are few CTCs. Therefore analyzing them is not easy work for pathologists. In this study, we develop a method which detects circulating tumor cells in fluorescence microscopy images automatically. Our proposed method has three steps. First, we extract cell regions in microscopy images by using filtering processing. Second, we separate the connecting cell regions into single cell regions,based on the branch and bound algorithm. Finally, we identify CTCs by using logical conjunction method. We demonstrated the effectiveness of our proposed method using 6 cases（5040 microscopy images）, and we evaluated the performance of CTCs identification. Our proposed method achieved, a true positive rate of 95.27 [%] and a false positive rate of 6.172 [%] respectively. And we confirmed the effectiveness of the logical conjinction for CTCs identicication.</p>

DOI： 10.11318/mii.34.151

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006267703

記述言語：日本語   掲載種別：研究論文（学術雑誌）

<p>Lung cancer is one of the most important cancer in the world. Among them, Ground Glass Opacity（GGO）has a hazy area of increased attenuation in the lung image. In recent years, development of a Computer Aided Diagnosis （CAD）system for reducing the burden on work load and improving the detection rate of lesions has been advanced. In this paper, we propose a CAD system to extract GGO from CT images. Firstly, we extract the lung region from the input CT images and remove the vessel, and bronchial region based on 3 D line filter algorithm. After that, we extract initial GGO regions using concentration and gradient information. Next, we calculate the statistical features on the segmented regions. After that, we classify GGO regions using support vector machine（SVM）. Finally, we detect the final GGO regions using deep convolutional neural network（DCNN）. The proposed method is tested on 31 cases of CT images from the Lung Image Database Consortium（LIDC）. The results demonstrate that the proposed method has 86.05[%] of true positive rate and 39.03[/case] of false positive number.</p>

DOI： 10.11318/mii.34.70

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006846732

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Singapore   Singapore   2016年11月22日  -  2016年11月25日

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2016 ICIC International. Rheumatoid arthritis and osteoporosis are two of the major diseases related to the phalanges. Diagnostic imaging is often used to diagnosis them. Especially, observing the temporal changes of the shape or internal structure of the phalanges in phalange CR images is an effective way to detect these diseases. However, there are various problems in image diagnosis, such as the evaluation of diagnosis is generally empiric, and the burden of reading images is heavy. In order to solve these problems, in this paper, we develop a computer aided diagnosis (CAD) system for automatically diagnosing, which includes segmentation of knuckles, registration of temporal images and features analysis of phalange regions. In the segmentation part, we perform the multi scale gradient vector flow (MSGVF) snakes to improve the precision of extracted results. Also we have developed an image registration technique based on salient region features (SRF) method. Two image features are used to train the artificial neural network (ANN) to identify the abnormal knuckle regions.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84991355197&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2016. This paper proposes a method of detecting prominent regions from an image using a global and local saliency measure. Detection of prominent regions from an image is important, because it is expected to improve the precision and the processing speed of object recognition through image processing. The proposed method employs two spatial redundancies, global saliency and local saliency. Global saliency is calculated by comparing the value of an interested pixel with the mode of the values of all pixels. On the other hand, local saliency is calculated using the relation among local pixels. In calculating the local saliency, the proposed method employs human visual characteristics, i.e., complementary color harmony to detect salient regions. Saliency map is made by integrating the global saliency and the local saliency considering some weights. Experimental results show the effectiveness of the proposed method.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84975702036&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2016 ICIC International. There are not a few visually impaired people in our society. When they want to go for a walk, it is difficult for them to perceive moving objects around them. When they walk on the sidewalk, they must pay strong attention to those coming closer to them or to moving objects. Therefore, it is indispensable for them to obtain the information on moving objects around them. This paper proposes a method of extracting moving objects, in particular, pedestrians, from self-wearable camera images. In the proposed system, a user is supposed to wear a camera and a PC and, by performing image analysis of the video taken from the camera, it acquires surrounding pedestrians’ information including their peculiar characteristics. The information is fed back to the user to realize his/her safe walk. Experimental results are shown and the performance of the proposed method is evaluated. The proposed system is primarily intended to be used by visually impaired people, but it may also be used by any pedestrian who is not paying much attention to his/her surroundings.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84959361584&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2016 ICIC International. In recent years, the equipment on walk support for visually impaired persons has spread to a certain extent. A studded paving block and a sound-type signal are set on the sidewalk or on the road in many places. However, there are still problems such that the studded paving block is still installed at limited places, or the sound of the signal is obscure because of the environmental noise such as roaring traffic or heavy rain. Therefore, a system that can support visually impaired persons in more effective way is necessary, such as a system which finds a signal or a traffic sign automatically and provides its information to the visually impaired. In order to realize such a system, this paper proposes a method of detecting pedestrian signals and crosswalk signs by a camera and a computer which we call MY VISION system. In the proposed method, color information is used in the first step to restrict the search for traffic signal and sign candidates, and then the HOG feature is introduced to describe the candidates by feature vectors. Recognition of the signals and the signs is performed by applying randomized trees to the candidates. The color sign of the pedestrian signal is also discriminated by using color information. Experimental results are shown and the method is evaluated.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84959357672&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2016年01月20日  -  2016年01月22日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2016年01月20日  -  2016年01月22日

記述言語：英語   掲載種別：研究論文（学術雑誌）

本稿では読影医師への負担軽減と読影精度の向上を目的とし，LIDCデータベース上の胸部CT画像からのGGO候補領域の自動抽出法を提案する．手法としては，まず肺野領域の抽出を行い，得られた肺野領域に対し，3D Line Filterによる血管・気管支領域の除去を行う．その後，濃度・勾配閾値処理により初期GGO候補領域を抽出する．そして初期GGO候補領域のセグメンテーションを行い，統計的特徴量を算出する．最後に，特徴量を基に識別器を構築し，最終的なGGO候補領域を決定する．本稿では，識別器としてニューラルネットワークとサポートベクターマシンの2種類を用い，両者の識別性能を比較する．実験では，提案法をLIDCデータベース上の胸部CT画像31症例に適用し，その結果に対する考察と有用性を述べる．

DOI： 10.3169/itej.70.J178

Scopus

CiNii Article

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84979211524&origin=inward

担当区分：責任著者   記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）

<p><i>Nowadays the population of elderly people who live alone has been growing continuously. When they encounter an accident at home, they won't be able to ask for help in the worst case. Under these circumstances, development of an automatic system to care elderly people living alone becomes increasingly important. We have been developing a computer vision system which finds abnormal motions of elderly people among their daily activities indoors. In this paper, we propose a new method of human motion representation called 3D-MHIs </i>(<i>3-dimensional-Motion History Image</i>)<i> based on FoE (Focus of Expansion) and a method of human motion recognition employing 3-D Hu moments by use of a single camera view. The method can recognize not only those motions parallel to a camera lens but also those motions toward depth direction. The effectiveness of the proposed method was experimentally shown.</i><i> </i></p>

DOI： 10.24466/pacbfsa.29.0_33

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130007979895

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2015 Institute of Control, Robotics and Systems - ICROS. In recent years, medical imaging is important in medical diagnosis for early detection of lesions. However, a large number of images increases the stress to the radiologist. Therefore, CAD (Computer Aided Diagnosis) system is expected to reduce such burden.

DOI： 10.1109/ICCAS.2015.7364602

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84966322344&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2015 Institute of Control, Robotics and Systems - ICROS. Recently, CT (Computed Tomography) scanner is used for detecting the abnormalities because CT scanner gradually becomes high resolution and high speed. However, with the improvement of the resolution of CT images, the number of CT images becomes huge. Therefore, radiologists have to analyze huge number of images and they sometimes misdiagnoses are happened. Hence, to deal with this problem the CAD (Computer Aided Diagnosis) system is developed. One of the CAD systems called temporal subtraction technique is useful to detect abnormalities in medical field. There is no viewer system which displays abnormal region using temporal subtraction technique. In this paper, we develop a novel temporal subtraction technique to help the radiologists to reduce diagnostic ti me and improve the diagnostic accuracy.

DOI： 10.1109/ICCAS.2015.7364601

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84966267281&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2015 Institute of Control, Robotics and Systems - ICROS. Recently, the necessity of environmental regulation, low fuel consumption, and natural energy development is proposed by environmental issues. So the demands of power transistor devices are increased. But measurement technique of the current distribution is not keeping up with further miniaturized and integrated were needed in present condition. Now, therefore, ensuring security attended high functionalization is a subject. IGBT (Insulated Gate Bipolar Transistor) is the device that used for wide range of power devices. So we are developing imaging system used non-contact sensor arrays aimed to IGBT production line. In this paper, we propose a development of a supporting system for visual inspection of IGBT device based on statistical feature and complex multi-resolution analysis. First, this performs signal de-noising after entering well-known good data and measured data. Second, the statistical feature is expressed the difference between good data and measured data are calculated. Last, classifying of good and inferiority is performed based on the result of threshold processing. In the paper, we applied our algorithm to 28 sample data including 20 good data and 8 inferiority data.

DOI： 10.1109/ICCAS.2015.7364603

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84966269121&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2015 Institute of Control, Robotics and Systems - ICROS. Kinect is always used as a device to estimate posture. However, there are difficult to estimate the posture in the case of using a Kinect only. Therefore, we propose a method to estimate more accurately posture by synthesizing the posture obtained by Kinect and estimated by the regression analysis. In the regression analysis, we associate the HOG features and joint parameters that consists of 20 coordinate points. Posture data used for learning of the regression model is used difficult posture be obtained with Kinect. Similarity in brightness between frames at around each joint of the skeleton obtained by regression analysis and Kinect is calculated. Then we synthesize the posture by calculating a weighted average. In addition, RVM regression model is used to improve the accuracy of representing the posture by the proposed method.

DOI： 10.1109/ICCAS.2015.7364600

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84966351140&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Indonesia   Surabaya   2015年09月29日  -  2015年09月30日

DOI： 10.1109/ELECSYM.2015.7380826

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Dalian   2015年08月20日  -  2015年08月22日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Dalian   2015年08月20日  -  2015年08月22日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Dalian   2015年08月20日  -  2015年08月22日

記述言語：英語   掲載種別：研究論文（学術雑誌）

近年,肺癌による死亡者数の増加から,病変部の早期発見・治療が重要視されている.そのため,胸部MDCT画像を用いた精密検査などが進められている.MDCT画像は,癌検出が容易である反面,読影を行う画像枚数が多く,医師への負担増が懸念されている.そこでコンピュータ支援診断システムによる,読影医師への負担軽減などが期待されている.このシステムの一つとして,経時的変化を強調する経時的差分像技術がある.この技術より得られた差分像から結節状候補陰影を検出するための研究が行われ,医師への診断支援の実現が期待されている.本稿では,胸部MDCT画像の経時的差分像より直径20[mm]以下の結節状候補陰影を抽出し,特徴量解析に基づく2段階AdaBoostによる病変候補陰影を識別するためのシステムの開発を行い,実画像による性能評価を行ったところ,96.8[%]の識別率を得た.

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/110009987542

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Hangzhou   2015年07月28日  -  2015年07月30日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Hangzhou   2015年07月28日  -  2015年07月30日

Kyutacar

記述言語：日本語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.1007/s10015-014-0195-4

Scopus

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2015. The motion sequences of human actions have its own discriminating profile that can be represented as a spatiotemporal template like Motion History Image (MHI). A histogram is a popular statistic to present the underlying information in a template. In this paper a histogram oriented action recognition method is presented. In the proposed method, we use the Directional Motion History Images (DMHI), their corresponding Local Binary Pattern (LBP) images and the Motion Energy Image (MEI) as spatiotemporal template. The intensity histogram is then extracted from those images which are concatenated together to form the feature vector for action representation. A linear combination of the histograms taken from DMHIs and LBP images is used in the experiment. We evaluated the performance of the proposed method along with some variants of it using the renowned KTH action dataset and found higher accuracies. The obtained results justify the superiority of the proposed method compared to other approaches for action recognition found in literature.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84957063027&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2015 SPIE. To obtain an effective interpretation of organic shape using statistical shape models (SSMs), the correspondence of the landmarks through all the training samples is the most challenging part in model building. In this study, a coarse-tofine groupwise correspondence method for 3-D polygonal surfaces is proposed. We manipulate a reference model in advance. Then all the training samples are mapped to a unified spherical parameter space. According to the positions of landmarks of the reference model, the candidate regions for correspondence are chosen. Finally we refine the perceptually correct correspondences between landmarks using particle filter algorithm, where the likelihood of local surface features are introduced as the criterion. The proposed method was performed on the correspondence of 9 cases of left lung training samples. Experimental results show the proposed method is flexible and under-constrained.

DOI： 10.1117/12.2179122

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84925431291&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Tainan   2015年01月11日  -  2015年01月13日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Ooita   2015年01月10日  -  2015年01月12日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Ooita   2015年01月10日  -  2015年01月12日

記述言語：英語   掲載種別：研究論文（学術雑誌）

関節リウマチは代表的な骨疾患で,病態が進行すると関節の変形や機能障害が生じ患者のQOL(quality of life)を著しく低下させる.これらの診断には画像診断が有効であるが,医師の主観的な評価による診断精度のばらつきや,画像枚数の増加に伴う医師への負担増加などの問題が現存するため,これらの問題を克服し医師の負担を軽減することが重要な課題である.手のCR画像から定量的な評価支援を行うためのコンピュータ支援診断システムの必要性に応えるため,本稿では,手のCR画像からの指骨領域の自動位置合わせ法を提案する.指骨領域は同一被験者の過去と現在の手のCR画像から指骨領域をそれぞれ自動抽出したものを用いる.位置合わせ法としては,指骨領域の関心領域上のエントロピーに基づく顕著領域特徴を求め,過去と現在画像上の顕著領域特徴間の関係性から最適な変形量を導出し,剛体変形を与えることにより,両画像の位置合わせを行い,経時的な変化部分を検出する.提案手法を3症例の過去と現在画像に適用し,良好な結果を得た.

DOI： 10.24466/jbfsa.17.2_35

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/110010047578

記述言語：英語   掲載種別：研究論文（学術雑誌）

手の骨の疾患,特に指骨の疾患における画像診断では,主に骨を直接撮影可能なX線やCTなどが用いられ,骨のすりへりや破壊の有無を診断する.しかし,MRIを用いた画像診断により,CTなどでは観察が困難な症状の診断に有効な場合があり,早期発見・診断が期待されている.このような背景から,近年の医療診断における画像診断の占める割合は向上しているが,画像診断では,医師による診断結果のばらつきや,画像枚数の増大による医師への負担増加が懸念されている.そのため,定量的な解析を行うための,コンピュータ支援診断システムが必要となる.本論文では,手のMR画像からの指骨の疾患の定量的な評価を行うための前段階としてMSGVF Snakesによる指骨領域のセグメンテーション手法を提案する.

DOI： 10.11485/itetr.39.43.0_11

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/110010015999

記述言語：英語   掲載種別：研究論文（学術雑誌）

In this paper, we develop a CAD (Computer Aided Diagnosis) system for assisting in visual screening. In the first stage, we calculate the statistical features such as density, shape and textures on segmented candidate regions. Finally, we implement a classifier based on support vector machine to distinguish final candidate regions. We applied the proposed method to 31 CT image sets in the Lung Image Database Consortium (LIDC) which is supplied by National Center Institute (NCI).

DOI： 10.11485/iteac.2015.0_31B-2

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006742266

記述言語：英語   掲載種別：研究論文（学術雑誌）

In the field for visual screening, it is difficult task to detect blood vessel regions with high accuracy detection rates since blood vessel regions are in contact with bone region on the CTA images. In this paper, we propose a new method for segmentation of arterial area, which are obtained CTA based on gradient features, and satisfactory experimental results are obtained. Some experimental results are shown with discussion.

DOI： 10.11485/iteac.2015.0_31B-1

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006742267

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Kitakyushu   2014年12月03日  -  2014年12月06日

DOI： 10.1109/SCIS-ISIS.2014.7044905

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Kitakyushu   2014年12月03日  -  2014年12月06日

DOI： 10.1109/SCIS-ISIS.2014.7044743

Kyutacar

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Kitakyushu   2014年12月03日  -  2014年12月06日

記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.2991/jrnal.2014.1.3.3

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

© 2014, ISAROB. This paper proposes a method of detecting moving objects using sequential inference of the background in a video taken with a moving camera. In the video taken using a moving camera, all positions of pixels change every frame. The positions of the background pixels in the image frame T are not the same as the positions of the background pixels in the image frame T + 1. 2D projective transform can be used to find changes in the pixel position every frame. Bilinear interpolation with four nearest pixels around the pixel in image frame T which corresponds to a pixel in the image frame T+1 can be used for creating a background model at T + 1. Having obtained the background model, a pixel in image frame T + 1 can be determined if it is a background pixel or a foreground pixel. The detection results of the proposed method are compared with the ground truth to determine the effectiveness of the proposed method.

DOI： 10.1007/s10015-014-0168-7

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84939889507&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

France   Paris   2014年10月27日  -  2014年10月30日

DOI： 10.1109/ICIP.2014.7025288

Scopus

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

CiNii Article

記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.2991/jrnal.2014.1.2.13

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Hakaido   2014年09月09日  -  2014年09月12日

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：日本語   掲載種別：研究論文（学術雑誌）

Kyutacar

CiNii Article

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（学術雑誌）

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 IEEE. A medical image fusion technique provides insight into lesion diagnosis and spatial relationships among anatomical structures. In radiotherapy planning for CyberKnife treatment, for instance, the accurate area being irradiated is determined and calculated based on a fusion image of CT and MR images. Hence the risk of the normal tissues being irradiated could be brought down. However, commonly used mouse-based registration yields variations in the accuracy of the results due to manual operations. In addition, there are some pending issues such as the increased physical burdens to the operators and the requirement for a substantial investment of time. Therefore, in recent years many researches have attempted to automate the registration for multimodality images. In this paper, a CAD (Computer Aided Diagnosis) system is developed to assist radiotherapy planning for CyberKnife treatment. We propose a new automatic image registration technique for image fusion of head CT (Computed Tomography) and MR (Magnetic resonance) images. In our method, the minimization of ratio image uniformity and the maximization of normalized mutual information on VOIs (Volumes Of Interest) are performed based on Levenberg-Marquardt algorithm. We evaluated our proposed method by 5 clinical cases and discussed the accuracy of the registration results with computational time.

DOI： 10.1109/SCIS-ISIS.2014.7044694

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84959398969&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 IEEE. Developing of computer aided diagnosis (CAD) system which analyzes the medical image by a computer and provides the doctor as a 'second opinion' the result is becoming increasingly popular. Temporal subtraction technique which is one of the CAD systems supports the observation by performing the differential operation of the current image and the previous one of the same patient and emphasizing temporal changes. Process of registration between images exist in this technique, this accuracy is greatly affects to the temporal difference image. However, registration method adaptable to a variety of images with high accuracy has not been established yet. So in this paper, we propose a three-dimensional non-rigid image registration method generic and high precision based on finite element method (FEM). In the proposed method, in order to improve the high cost is the biggest problem of the FEM, introducing the feature point detection in generating a tetrahedral mesh model. Furthermore, we introduce global matching to achieve an improvement of the execution efficiency. In this paper, we verify the effectiveness by the result applying the model data and 3D thoracic CT Images the proposed method.

DOI： 10.1109/SCIS-ISIS.2014.7044691

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84946530780&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 IEEE. In recent years, technique that estimates the movement of people from images has been studied actively. It is a technique to recognize and understand the intentions of human behaviors, using the results of sensing the state of a person by the time. It is widely applied for example to recognize gesture command of user or to detect suspicious persons using surveillance camera. Moreover, recent studies using Kinect is thriving. Kinect is a device that performs tracking and posture estimation of persons. However, the system has some problems. One of them is that Kinect's tracking system is vulnerable to occlusion. It is necessary to be improved. In this paper, we perform tracking of a person using a particle filter algorithm. We use a feature quantity obtained by combining the color information and depth information as the criteria of the particle filtering. We perform experiments at two situations. And, the results of the proposed method are compared with the results of the conventional method [6] . In certain circumstances, the results indicate that our proposed method is more accurate than the conventional method.

DOI： 10.1109/SCIS-ISIS.2014.7044772

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84946531166&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 IEEE. Image registration is an important and a fundamental task in computer vision and image processing field. For example, to make a surgical plan for head operation, the surgeons should gain more detailed information from CT angiography (CTA) and MR angiography (MRA) images. And the abnormalities can be easily detected from the fusion image which is obtained from two different modalities. One of the multiple modal image registration methods is matching the CTA and MRA, by which the image of head vascular could be enhanced. In general, the procedure for fusion is completed manually. It is time-consuming and subjective. Particularly the anatomical knowledge is required as well. Therefore, the development of automatic registration methods is expected in medical fields. In this paper, we propose a method for high accurate registration, which concentrates the structure of head vascular. We use 2-D projection images and restrict volume of interests to improve the processing affection. In experiments, we performed our proposed method for registration on five sets of CTA and MRA images and a better result from our previous method is obtained.

DOI： 10.1109/SCIS-ISIS.2014.7044848

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84988302681&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 IEEE. Lung cancer is the most common cause of death from cancer worldwide. Therefore, for the purpose of early detection of cancer, mass screening and thorough examination have been carried out. Lung cancer is detected easily by using chest MDCT (Multi Detector-row Computed Tomography) images. However, radiologists are apprehended burden by many chest MDCT images which are required interpretation of radiograms. So the CAD (Computer Aided Diagnosis) systems that could relieve radiologist's stress and diagnose accuracy could be improved are expected. One of the CAD systems, temporal subtraction technique that emphasized time-dependent change is reported. This technique is used for diagnosis assistance of detected candidate nodules from CT images. In this paper, the candidate nodules under 20[mm] are extracted from temporal subtraction images. We highlighted the candidate nodules based on features analysis of images. We applied proposed method to 31 cases of chest MDCT images in which the number of nodules was more than one. We got a result of TPR:96.9[%], FPR:6.45[/case].

DOI： 10.1109/SCIS-ISIS.2014.7044780

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84946531131&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 IEEE. Detection of pulmonary nodules with ground glass opacity (GGO) is a difficult task in radiology. Follow up is often required in medical fields. But diagnosis based on CT images are dependent on ability and experience of radiologists. In addition to that, enormous number of images increase their burden. So, to improve the detection accuracy and to reduce the burden of doctors, a CAD (Computer Aided Diagnosis) system is expected. So, in this paper, we propose an automatic algorithm for GGO detection on CT images. At first, vessel areas are removed from original CT images by using 3D Line Filter and then candidate regions are detected by threshold processing. After that, we calculate statistical features of segmented candidate regions and use artificial neural network (ANN) to distinguish final candidate regions. We applied the proposed method to 31 CT image sets in the Lung Image Database Consortium (LIDC) which is supplied by National Center Institute (NCI). In this paper, we show the experimental results and give discussions.

DOI： 10.1109/SCIS-ISIS.2014.7044692

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84946531517&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

Traffic accidents are decreasing due to the influence of technology advancement; however accidents still occur due to carelessness of drivers. Therefore, many researchers have been studying how to realize an advanced safety system. The Histograms of Oriented Gradients (HOG) feature is well known as a useful method of detecting a standing human in various kinds of backgrounds. Unlike a person, a bicycle can appear differently from various angles. In this paper, we propose a method of detecting a bicycle on the road using improved HOG feature named MSC-HOG feature and the Real-AdaBoost algorithm. Experimental results and evaluation show satisfactory performance of the proposed method. © 2014 ISSN 1349-4198.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84893415119&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

© 2014 Institute of Control, Robotics and Systems (ICROS). Rheumatoid arthritis and osteoporosis are two common orthopedic diseases. Rheumatoid arthritis is a disease that inflammation occurs in the joint, which always causes the joints are able to move freely. Osteoporosis is a disease that bone mineral content is reduced and risk of fragility fracture increases. As one of the diagnostic methods, medical imaging by photographed CR equipment has been widely accepted. However, some problems such as mass screening data sets and mis-diagnosis are still remained in visual screening. In order to solve these problems and reduce the burden to physicians, needs of an automatic diagnosis system capable of performing quantitative analysis is anticipated. In this paper, we carry out the development of a segmentation method of phalanges regions from CR images of the hand to perform a quantitative evaluation of rheumatoid arthritis and osteoporosis. The proposed method is carried out crude segmentation of phalanges regions from CR images of the hand, and extracts the detailed phalanges regions by Multi Scale Gradient Vector Flow Snakes (MSGVF) method. In our study, we performed Snakes algorithm to give an initial control points on MSGVF algorithm. We applied our method on three pairs of CR temporal images of phalanges regions, which are called as the previous images and the current images. We got the segmentation results of 5.95 [%] of false-positive rate and 92.9 [%] of true-positive rate.

DOI： 10.1109/ICCAS.2014.6987755

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84920195728&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2014年01月22日  -  2014年01月24日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2014年01月22日  -  2014年01月24日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Beppu   2014年01月22日  -  2014年01月24日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Ooita   2014年01月11日  -  2014年01月13日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Ooita   2014年01月11日  -  2014年01月13日

Kyutacar

記述言語：日本語   掲載種別：研究論文（学術雑誌）

DOI： 10.2534/jjasnaoe.18.115

Kyutacar

CiNii Article

記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.1587/transinf.E96.D.2882

Scopus

CiNii Article

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Kobe   2013年12月15日  -  2013年12月17日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Kobe   2013年12月15日  -  2013年12月17日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Naha   2013年11月05日  -  2013年11月08日

DOI： 10.1109/ACPR.2013.174

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Since it is difficult to choose which computer calculated features are effective to predict the malignancy of pulmonary nodules, in this study, we add a semantic-level of Artificial Neural Networks (ANNs) structure to improve intuition of features selection. The works of this study include two: 1) seeking the relationships between computer-calculated features and medical semantic concepts which could be understood by human; 2) providing an objective assessment method to predict the malignancy from semantic characteristics. We used 60 thoracic CT scans collected from the Lung Image Database Consortium (LIDC) database, in which the suspicious lesions had been delineated and annotated by 4 radiologists independently. Corresponding to the two works of this study, correlation analysis experiment and agreement experiment were performed separately © 2013 IEEE.

DOI： 10.1109/EMBC.2013.6610786

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84886486763&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Nagoya   2013年09月14日  -  2013年09月17日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Nagoya   2013年09月14日  -  2013年09月17日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Nagoya   2013年09月14日  -  2013年09月17日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Nagoya   2013年09月14日  -  2013年09月17日

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

UK   York   2013年08月27日  -  2013年08月29日

DOI： 10.1007/978-3-642-40246-3_41

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Kyoto   2013年05月21日  -  2013年05月23日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Points Distribution Model (PDM) has been established as one of the robust tool for medical image segmentation and shape analysis. A main challenge to apply PDM is model correspondence across a set of training samples primarily. In this study, we propose a direct correspondence method based on spherical conformal mapping theory. Firstly, map the genus zero surface to a spherical domain. Constrain a group of Möbius transformations by three selected datum points on surface. In this way, the attitudes of all training samples are corresponded. Then place the landmarks across the whole training set automatically according to the reference model. To require the statistic data, the distribution of landmarks should be accounted in a uniform framework. We align these three datum points using an analytic method. Our proposed method was performed on the correspondence of 4 cases of left lung training samples. Compared with the manual correspondence results, the mean error of Euclidean distance for each landmark was 12mm. © 2013 Springer-Verlag.

DOI： 10.1007/978-3-642-29305-4_247

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84876020662&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Human action recognition from video clips has become an active research field in recent years. Each action has its unique shape and a motion sequence can be suitably represented by a histogram. In this paper a histogram based action recognition method is presented. Motion history images are a good spatiotemporal template for action representation. In the present method, we use local binary patterns of directional motion history images for the histogram representation. We measured the performance of the proposed method along with some variants of it by employing KTH action dataset and found higher accuracy. The presented results also justify the superiority of the proposed method compared to other approaches for action recognition found in literature. © 2013 IEEE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84902492748&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper describes a dental trainings imulation sys tem with a jaw mode l using a computer vis ion technique for scaling and root planing (SRP) . The SRP is a cornerstone of treatment of periodontal diseases. We have difficulty in watching the movement of the working end of a scaler in a mouth. We propose a method of recovering the movement of the scaler employing a marker attached to the scaler. The performance of the proposed method is shown experimentally.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84888617888&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Side-scan and forward-looking sonars are some of the most widely used imaging systems for obtaining large scale images of the seafloor, and their use continues to expand rapidly with their increased deployment on autonomous underwater vehicles. However, it is difficult to extract quantitative information from the images generated from these processes, particularly for the detection and extraction of information on the objects within these images. We propose in this paper an algorithm for automatic detection of underwater objects in side-scan images based on machine learning employing adaptive boosting. Experimental results show that the method produces consistent maps of the seafloor.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84888617459&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a technique to separate foreground objects from the background in a video taken with a moving camera. In this technique, the first image frame is considered as the first background. The next image frame is an image containing foreground objects. The next image frame is conditioned to have the same position with the first background. Adjustment of the position between the two image frames is done by determining the image features of the two images using Harris corner detector method. After image features have been obtained, feature point correspondence between the first background image and a subsequent image is searched for using Lucas-Kanade tracker. Outlier pairs are discarded by RANSAC. By the employment of a set of feature point pairs, the 2D projective transform is computed between the two images. Furthermore it is determined whether a pixel is included in the foreground or the background. Experimental results show satisfactory performance of the proposed method.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84888592154&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Obstacles detection is an important technology in autonomous collision avoidance systems. We have already proposed an obstacle detection method based on background modeling. But this method detects 2D and 3D objects simultaneously. Since these 2D objects are not dangerous to driving, they will reduce the accuracy of detection if they are detected as obstacles. In order not to detect these 2D objects, this paper proposes a method for classifying 2D objects and 3D objects. The proposed method first estimates the camera motion parameters from the correspondences of feature points between two successive images. We calculate the 3D positions of the feature points on a detected object in the world coordinate system using triangulation. Then we estimate the parameters of the road plane using 3D positions of those feature points. Finally we calculate the distances from the 3D positions of the feature points to the road plane. Based on these distances, we can classify 2D objects and 3D objects. Experimental results show satisfactory performance of the proposed method.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84888611260&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Understanding the distribution of seafloor sediment using a side-scan sonar is very important to grasp the distribution of seabed resources. This task is traditionally carried out by a skilled human operator. However, with the appearance of Autonomous Underwater Vehicles, automated processing is now needed to tackle the large amount of data produced and to enable on the fly adaptation of the missions and near real time update of the operator. We propose in this paper a method that applies a subspace method and higher-order local auto-correlation feature to the acoustic image provided by the side-scan sonar to classify seabed sediment automatically. In texture classification, the proposed method outperformed other methods such as gray level co-occurrence matrix and Local Binary Pattern operator. Experimental results show that the proposed method produces a consistent map of a seafloor. © 2013 IEEE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84902467004&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

We present a motion classification approach to detect movements of interest (abnormal motion) based on optical flow. By tracking all feature points of a moving human in successive frames, we calculate the coordinate space and create feature space. This is done directly from the intensity information without explicitly computing the underlying motions. It requires no foreground segmentation, no prior learning of activities, no motion recognition and no object detection. First, we determine the abnormal scene and speed by using the velocity histogram. Then by using k-means clustering over velocity orientation and magnitude, we determine the abnormal direction. The performance of the proposed method is experimentally shown.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84888608440&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

Place-march of corresponding landmarks is one of the major factors influencing 3D Points Distribution Model (PDM) quality. In this study, we propose a semi-automatic correspondence method based on surface parameterization theory. All the training sets are mapped into a spherical domain previously. The rotation transformation of training samples is regarded as spherical rotation of their maps. We solve it by comparing the density distribution of surface map of training sample with respect to the reference model. Simultaneously, the corresponding landmarks across the whole training set are marketed depending on the spherical coordinates on parameter domain. In this paper, we also compared the corresponding results with two constraint conditions of spherical conformal mapping: 3 datum points constrain and zero-mass constrain. Experimental results are given for left lung training sets of 3D shapes. The mean result with the 3 datum points constraint and the zero mass-center constraint was 21.65 mm and 20.19 mm respectively.

DOI： 10.20965/jaciii.2013.p0018

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84879455289&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

Recently, the arteries sclerosis obliterans (ASO) or called peripheral arterial disease (PAD) typically caused by chronic ischemia of limbs increases remarkably. As one of the diagnosis methods, the image diagnosis methods such as MR image are applied in medical fields. In this paper, we propose a vascular extraction method using fresh blood imaging (FBI) method, as well as apply it to computer aided diagnosis (CAD) system. Especially, to prevent the spread outside of the region and improve the segment accuracy of peripheral artery areas, we introduce particle filter algorithms. We performed our method on automatic artery regions detection using non-enhanced MR images. Furthermore, we compared the extracted results to gold standard data and analyzed accuracy by receiver operating characteristic (ROC). The effectiveness of our proposed method and satisfactory of its detected accuracy were confirmed.

DOI： 10.20965/jaciii.2013.p0318

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84879329594&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, the death rate due to the liver cancer rises remarkably. To reduce the rate, the early detection of the disease is important. To detect diseases in early stage which are concern cancer, image diagnosis such as CT image is used in medical fields. On the other hand, the burden to a radiologist becomes increase. Therefore, the development of a system reducing the burden of the radiologist is important. In order to diagnose abnormalities based on medical imaging there are some reports. But, there is no report which is concern with detecting abnormality on liver disease based on temporal subtraction technique for abdominal CT image. As one of the methods to analyze abnormalities on visual screening, temporal subtraction technique is useful. This technique subtracts past image to current one. To obtain the good performance based temporal subtraction technique, image registration is most important task. In this paper, we propose a registration method for liver CT image using voxel matching techniques. We describe our registration method from two CT image which obtained deference time series and shows experimental results with discussion. © 2012 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872554258&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In this paper, we propose a new surface-based method to solve the rotation transformation of 3D rigid registration. Different from the criterion that minimize the sum of the squares of the errors between the corresponding parts of objects, here we estimate the spatial modalities of objects by their orientations. In details, Gauss map of surface is introduced to measure the distribution of surface curvature. To depict the characters of surface, we add sign weights to Gauss sphere according to the local total curvature. The stereographic projection is utilized to transform the spherical statistics to bifacial plane. As well, an optimization solution is done by iterated algorithm. We applied our proposed method on lung triangular surface registration. The experiments were completed by mutual information of block statistic. © 2012 IEEE.

DOI： 10.1109/ICSMC.2012.6377933

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872393612&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, radiologists can easily screening with high accuracy based on CTA (Computed Tomography Angiography) images and/or MRI (Magnetic Resonance Imaging) which is supported by medical imaging technology. On the other hand, visual screening makes burden to doctors which cause increasing the images and limited times. To avoid these problems, CAD (Computer Aided Diagnosis or Detection) system is developed by reducing the burdens and to improve the diagnostic accuracy. Problems of diagnosis on legs from CTA are burdens to doctors and detection of blood vessel by manual. In the field for visual screening, it is difficult task to detect blood vessel regions with high accuracy detection rates since blood vessel regions are in contact with bone region on the CTA images. In this paper, we propose a new method for segmentation of arterial area, which are obtained CTA based on image processing by concentration gradients, and satisfactory experimental results are obtained. Some experimental results are shown with discussion. © 2012 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872535204&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Traffic accidents are decreasing under the influence of technology advancement. But the problems still remain that accidents occur due to carelessness of drivers. Therefore many researchers have been still studying to realize an advanced safety system. The Histograms of Oriented Gradients (HOG) feature is well known as a useful method of detecting a standing human in various kinds of the background. Unlike a human, a bicycle changes its appearance variously according to viewpoints. Hence, it is more difficult than detecting a human. In this paper, we propose a method of detecting a human on a bicycle using the Multiple-size Cell HOG (MSC-HOG) feature and the RealAdaboost algorithm. Experimental results and evaluation show satisfactory performance of the proposed method. © 2012 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872579239&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

It is necessary for a blind person to know the locations of the pedestrians or moving objects around him/her on a walk road for his/her safe walking. This paper proposes a system for detecting pedestrians or moving objects in front of a person on a walk road by a camera mounted on his/her body. The system judges if a moving object in front of the person is going to hit him/her in the near future by analyzing the motion vectors acquired on the camera images. In case there is a fear of hit, the system tells it to the person by sound. In this way, the proposed system helps safe walking of a blind person. Some experimental results are shown with discussion. © 2012 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872511199&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a novel technique for 3-D recovery of a non-rigid object, such as a human in motion, from a single camera view. To achieve the 3-D recovery, the proposed technique performs segmentation of an object under deformation into respective parts which are regarded as rigid. For high accuracy segmentation, multi-stage learning and local subspace affinity are employed in this stage. Each part recovers its 3-D shape by applying the factorization method to it. This is the initial solution. The shape of each part is then refined by applying a quadratic model to the initial solution. The entire 3-D recovery of the object is finally performed using the common points among the segmented parts. The experiments employing a synthetic non-rigid object and real human motion data show effectiveness of the proposed technique. © 2012 ICPR Org Committee.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84874574364&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Temporal Subtraction Technique is one of effective tools for diagnosis of lung cancer from thoracic CT images. By comparing two images of the same subject but shot at different time, the detection of temporal changing becomes facilitated. To acquire a more accurate subtraction image, the registration of these two images is critical. However it is not easy as the influence of the slight structures such as lung blood vessel. In this paper, a novel nonrigid image registration method based on vessel structure information is proposed. The similarity of the vessels structure is defined by means of the likelihood function of vessels structure and their direction. We combine this similarity of vessels with the intensity information of images. And the metrics are used as similarity measure in registration procedure. The proposed method has been applied to thoracic MDCT images, and the improvement of registration accuracy was investigated. The efficiency of our proposed method was indicated. © 2012 IEEE.

DOI： 10.1109/ICSMC.2012.6377932

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872395110&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, thorax MDCT images are used in visual screening for early detection of lung nodules. Radiologists can easily detect lung nodules on images, but it has enormous images and load of radiologist for visual screening. To reduce the load of radiologist and improve the detection accuracy, a CAD (Computer Aided Diagnosis) system is expected from medical fields. In the medical image processing fields, some related works are reported to develop the CAD system including temporal subtraction technique as helpful technical issues. In this paper, we propose a classification of lung nodules on temporal subtraction image based on image processing technique. At first, the candidate regions including nodules are detected by the multiple threshold technique in terms of the pixel value on the temporal subtraction images. Then, we remove vessel regions on nodules by the most suitable threshold technique and watershed method. Also we remove the false positives which are caused by mis-registration using selective enhancement filter, rule-base method and artificial neural networks. In this paper, we illustrate some experimental result which applied our algorithm to 31 chest MDCT cases including lung nodules. © 2012 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872550943&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

A temporal subtraction technique which is subtracted from previous image to current one is introduced as powerful tools in medical fields to diagnose abnormalities. It provided a computer aided diagnosis (CAD) tools on visual screening. Radiologist can detect lesions on image by compare the two images. It is because the subtraction image can enhance the temporal changes, such as shaped of new lesions and/or the temporal changes in existing abnormalities by removing most of the normal background structures by subtraction of a previous image from a current one. There are some technical reports to register the different images until now. But subtraction artifacts are still remained which are caused by mis-registration. In this paper, we propose a new method for temporal subtraction method on thoracic MDCT images using non-rigid image warping techniques based on free form deformation (FFD). We applied our method to two clinical cases of chest CT image sets and compare to conventional methods in terms of computational cost and accuracy. © 2012 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84872541007&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

The fundamental step to get a Statistical Shape Model (SSM) is to align all the training samples to the same spatial modality. In this paper, we propose a new 3D alignment method using surface parameterization theory to solve the rotation transformation of 3D rigid registration. It is a feature based alignment method which matches two models depending on comparing the distribution of spherical conformal map of vertices. Moreover, the stereographic projection is utilized to transform the spherical statistics to bifacial plane. The optimal solution is obtained by an iterated algorithm. We tested the rigid registration of left lung training samples. The availability of our proposed method was confirmed. © 2012 IEEE.

DOI： 10.1109/ICIEA.2012.6360834

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84871706167&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper presents a contour segmentation of cardiac region on ultrasound images viewed as a tracking framework. We develop a tracking model for contour segmentation and estimate the image contour using particle filter algorithm. The tracking model is initialized by projecting equip spaced radii from center point of the contour to the tracking boundary. The motion of each particle on the radii is governed by a system model. The samples likelihood is measured based on the gradient intensity or edge of the contour. Then the estimated contour is measured based on the mean estimate of the samples likelihood. Our proposed method is implemented to the ultrasound images of a cardiac and the satisfactory results are achieved. © 2012 IADIS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84887293035&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Beijing   2012年12月12日  -  2012年12月16日

DOI： 10.1109/ICCVE.2012.59

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Malaysia   Kuching   2012年11月21日  -  2012年11月24日

Kyutacar

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Malaysia   Kuching   2012年11月21日  -  2012年11月24日

記述言語：英語   掲載種別：研究論文（学術雑誌）

© ISAROB 2012. Peripheral arterial disease (PAD) is one of the reasons caused to the lower extremity atherosclerotic disease. Its diagnosis is needed to obtain much kind of the information of vascular morphology as well as the blood flow information based on hemodynamics. The diagnosis of the PAD using magnetic resonance imaging (MRI) equipment without contrast medium is available as a useful visual screening in clinical practice. In this paper, we propose a novel method for visualizing hemodynamics to arterial images obtained by a non-contrast enhanced magnetic resonance angiography (MRA) based on the Lucas–Kanade optical flow with the image pyramid processing, and satisfied experimental results are obtained.

DOI： 10.1007/s10015-012-0022-8

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84863670045&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2012年10月19日  -  2012年10月21日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2012年10月19日  -  2012年10月21日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

USA   Florida   2012年09月30日  -  2012年10月03日

DOI： 10.1109/ICIP.2012.6466898

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Akita   2012年08月21日  -  2012年08月23日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Akita   2012年08月21日  -  2012年08月23日

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Oita   2012年01月19日  -  2012年01月20日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a technique for recognizing Japanese finger-spelling using a sign language user's chest-mounted camera. Unlike existent systems, the technique employs a chest-mounted camera attached to a sign language user himself/herself and recognizes his/her sign language through the captured images of Japanese finger-spelling. We use a hand area picture of his/her hand and the MHIs (Motion History Images) for the Japanese finger-spelling recognition. For the recognition method, we employ the ICA (Independent Component Analysis). Furthermore, in order to recognize the Japanese finger-spelling from an animation (an image sequence), a character segmentation technique is also proposed. The performance of the proposed system is shown experimentally. © 2012 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84869417788&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a method of motion recognition in real night environment employing an infrared radiation camera. We record a video employing an infrared radiation camera to perform motion recognition at night. We use Directional Motion History Images (DMHIs) and Directional Motion Energy Images (DMEIs) for human motion representation. For the recognition of the motions, we employ the eigenspace method. In the experiment, it was confirmed that the human motion recognition at night is possible. © 2012 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84869392576&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Oita   Beppu   2011年01月19日  -  2012年01月21日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Oita   Beppu   2011年01月19日  -  2012年01月21日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a method of detecting obstacles from a video taken by a moving camera mounted on a vehicle by background subtraction. The background subtraction is often used to detect moving objects when camera is static. But according to the characteristics of a road, we can also employ Gaussian mixture model to detect all objects (either static or moving objects) on the road in the case of moving camera. Then we use two consecutive image frames, and warps the first image according to the geometrical relationship between these two images. The road area is then extracted by comparing the warped image with the second image. Using this road area, we can delete all things which are not obstacles. In the performed experiments, it is shown that the proposed method is able to detect obstacles such as vehicles and pedestrians on a road. © 2011 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866717620&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, there are many traffic accidents in turning right at an intersection. They are mainly caused by a driver's oversight of pedestrians and motorcycles that are occluded by oncoming cars. Therefore a system is necessary to detect moving objects such as oncoming cars and pedestrians at an intersection, and warn a vehicle driver. This paper describes a technique for detecting moving objects in turning right at an intersection when vehicle is stopping. Moving objects are detected by Mixture of Gaussians (MoG). In addition, we distinguish cars from pedestrians using the difference of the area size and the aspect ratio of detected objects. The object which is classified as a pedestrian is tracked using Lucas-Kanade Tracker. If the detected cars and pedestrians overlap or a car completely obscures pedestrians, we perform the estimation of pedestrian's location by using the information on past frames. By doing this, it is possible to detect pedestrians that drivers are actually difficult to see. The performance of the proposed technique was examined employing car videos and satisfactory results were obtained. © 2011 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84863041517&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

A traffic accident occurs mainly because of carelessness of a driver. To prevent such an accident, many systems which detect pedestrians from vehicle are proposed. But most of these systems give a warning to a driver every time when a pedestrian is detected. The effect of warning becomes lower if the frequency of warning increases in a busy area. The objective of this study is to develop a technique for detecting pedestrians from a vehicle and to arrange them according to their potential risks by analyzing their behaviors. In this paper, we propose a technique for segmenting pedestrians employing stereo camera images to analyze a pedestrian's behavior. First, foreground regions are extracted employing background estimation using the mixture of Gaussian model. After extracting foreground regions in an input image, these regions are segmented by using the distance information. By using this method, each shape of the detected pedestrians can be obtained even when pedestrians are overlapping with each other. The proposed technique was examined experimentally employing real video images and satisfactory results were obtained. © 2011 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84863049839&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, the number of accidents that pedestrians have law violation is in the tendency of decrease in Japan. However, accidents caused by pedestrians crossing a crosswalk or dashing into a crosswalk still have high ratio, and both accident sources account for 15% of the whole number of accidents caused by a pedestrian. Although many researches in ITS in which pedestrians are detected from in-vehicle cameras have been actively done to solve these problems, they usually employ standard cameras, and those pedestrians who exist outside of the camera view cannot be detected. In this paper, we employ a wide-angle camera which has wider view than a general camera and propose a technique for detecting pedestrians from the wide-angle image. Since, in a wide-angle camera image, every object becomes smaller, we propose a technique for detecting pedestrians employing optical flows converging to a FOE (Focus of Expansion). Experimental results show satisfactory performance of the technique. © 2011 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84863031172&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Studies on car vision have currently been practiced around recognizing a human enthusiastically. The Histograms of Oriented Gradients (HOG) feature has been proposed as useful feature for recognizing a human standing in various kinds of background. On the other hand, although a bicycle is important transportation vehicle in urban environment, its automatic recognition or detection is not an easy task for a computer vision system, because bicycle's appearance can change dramatically according to viewpoints and a person riding on the bicycle is a non-rigid object. Thus, automatic bicycle detection is an important research subject in an intelligent perception system using car vision. In this paper, we propose a method of detecting a bicycle and its driving direction using the HOG feature and RealAdaboost algorithm. Experimental results show satisfactory performance of the proposed method. © 2011 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866651335&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Automatic detection of human flows on a road by a computer vision system is of great importance mainly in surveillance systems, where human flows are observed by a camera and a computer analyzes the videos that the camera provides to detect a person having a different flow of movement, such as a person walking toward a certain direction while most of the people walk in the opposite direction, or a person running in a group of walking people. This paper describes a technique for finding a person having a different behavior or motion from others. The idea of the paper is to classify motion flows (or optical flows) extracted from a video into respective groups having respective directions of the motion by analyzing the motion flows. Experimental results show effectiveness of the proposed technique. © 2011 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866643383&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Detection of a human on a bicycle is an important research subject in an advanced safety vehicle driving system to decrease traffic accidents. The Histograms of Oriented Gradients (HOG) feature has been proposed as useful feature for detecting a standing human in various kinds of background. So, many researchers use currently the HOG feature to detect a human. Detecting a human on a bicycle is more difficult than detecting a human because a bicycle's appearance can change dramatically according to viewpoints. In this paper, we propose a method of detecting a human on a bicycle using HOG feature and RealAdaboost algorithm. When detecting a human on a bicycle, occlusion is a cause of decreasing detection efficiency. Occlusion is a serious matter in car vision research because there are occlusions in real transportation environment. We decide the next position of a human on a bicycle using object tracking. Experimental results and evaluation show satisfactory performance of the proposed method. © 2011 ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84856534795&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In the medical image processing field, segmentation from the CT image is one of the most important problems for analyzing the abnormalities and diagnosis on visual screening. Many related segmentation techniques have been developed for automatic extraction of ROI. It is however, there are still no fully automatic segmentation methods that are generally applicable to ROI based on CT image set. In this paper, we present a technique for automatic extraction of liver region on the MDCT images employing automatic construction of tree-structural image transformation (ACTIT). We propose a new technique for extraction of organs employing ACTIT with non-contrast and contrast image set in order to introduce temporal change information. We apply the proposed technique to three abdominal image set and satisfactory segmentation results are achieved. © 2011 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866687359&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Australia   Melbourne   2011年11月07日  -  2011年11月10日

DOI： 10.1109/IECON.2011.6120002

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Nagasaki   2011年10月31日  -  2011年11月

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Gyeonggi-do   2011年10月26日  -  2011年10月29日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Gyeonggi-do   2011年10月26日  -  2011年10月29日

Student Paper Award Finalist

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Gyeonggi-do   2011年10月26日  -  2011年10月29日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Gyeonggi-do   2011年10月26日  -  2011年10月29日

記述言語：日本語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月13日  -  2011年09月18日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月13日  -  2011年09月18日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月13日  -  2011年09月18日

2011年度計測自動制御学会 学術奨励賞技術奨励賞

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月13日  -  2011年09月18日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月13日  -  2011年09月18日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月13日  -  2011年09月18日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Japan   Tokyo   2011年09月05日  -  2011年09月09日

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

2011年09月28日  -  2011年10月01日

記述言語：日本語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：日本語   掲載種別：研究論文（学術雑誌）

記述言語：日本語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taiwan   2011年04月16日  -  2011年04月17日

記述言語：英語   掲載種別：研究論文（学術雑誌）

Image registration is one of the important techniques for diagnosis or treat- ment using medical images. Especially, it is very efficient to detect abnormalities by comparing two or more images. The images, which are provided from multi modalities or taken several different times, are used for diagnosis. One of the multi modal images is CT and MR imaging of the head for diagnosis and surgical planning indicates that physicians and surgeons gain important information from these modalities. In general, to simulate before the surgical operation, manual registration techniques are mainly per- formed on MR and CT images of the head. However, manual registration of the object may require several hours for analysis based on anatomical knowledge. In this paper, we describe a new method for image registration technique of head images obtained from CT and MRI. Our new method uses an optimal retrieval on neighbor volumes of interest (VOIs) and maximization of mutual information. This method also can reduce the pro-cessing time without decreasing the registration accuracy. © 2011 ISSN 2185-2766.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79251503899&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper describes a system for 3-D modeling of dynamic remote environments employing the images provided from cell-phone cameras and a communication network. In the system, we employ two cell-phones to acquire images. Each image is sent to a lab by an e-mail and employed for the 3-D modeling. For the algorithm of 3-D recovery, we employ a mobile stereo vision method, since this method is based on the factorization method and a steady solution can be expected. In the experiment, we tested the influence of the restoration error by using two cell-phones and obtained an excellent result. © 2011 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=81255179892&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recognition of the actions of several persons in a scene is gaining importance for human action analysis. Such a kind of system capable of recognizing human actions in a scene is proposed in this paper. We have adopted a fast human detector using Histograms of Oriented Gradients (HOG) and Support Vector Machine (SVM) to robustly trace individual persons. Person-wise bounding blobs are extracted and individual features are tracked in subsequent frames. The recognition is performed individually by comparing with the known motion templates using a high-speed action database. The technique guarantees improved performance in terms of detection and recognition. © 2011 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=81255147356&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper presents an improved spatio-temporal (XYT) approach for local interest point-based global action representation, considering the history of moving points in an action. The presented spatio-temporal representation demonstrate robust results and we compare the developed method with previous other method. This is a SURF-based method where we extract visual features to select candidate points based on the SURF detector. Afterwards, motion features are extracted by exploiting the local interest points and by employing optical flow. RANSAC is employed to reduce the unwanted outliers and improve the performance of the method. Based on an outdoor action dataset, we have found that the developed method demonstrate satisfactory recognition results. © 2011 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=81255210497&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

The structure from Motion (SfM) algorithm is an established method of shape recovery from a single video sequence. However, limitation of the method is the accuracy of the recovery due to poor texture of a recovered plane or a large scale object. We propose a technique for improving the precision of the recovery by applying the Manhattan-world constraint to the SfM algorithm, which assumes that the buildings are composed of vertical and horizontal planes. We show its effectiveness by the experiments performed in a real-life environment. © 2011 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=81255178501&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Human action understanding and recognition have various demands for different applications in the field of computer vision and human-machine interaction. Due to these issues, more than a decade, extensive researches are going on in this arena - to recognize various actions and activities. Researchers have been exploiting various action datasets and some of them become prominent. Though there are some good datasets, unfortunately, to have a strong survey on these datasets has been a long due. This paper attempts this and presents the key datasets and analyzes them in different perspectives. © 2011 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=81255147357&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a technique for 3-D recovery of a non-rigid object, such as a moving person, from a single camera view. Recovery of a non-rigid object is not possible from a single camera view without any condition. In this paper, we propose a single camera technique for recovering a non-rigid object under the condition that the object is composed of a set of rigid objects. The experiments employing real motion data show effectiveness of the proposed technique. © 2011 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=81255128065&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Bangladesh   Dhaka   2010年12月18日  -  2010年12月20日

主要論文集（会議）

DOI： 10.1109/ICELCE.2010.5700789

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Singapore   Singapore   2010年12月05日  -  2010年12月08日

主要論文集（会議）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

The development of an automatic human motion recognition system leads to the solution to the problems concerning the video-based applications in recognizing human activities. Such a system is to be investigated in the context of human motion analysis. Although there were a large number of researches in this area for a long time, there was little attention given to the development of a structured database for successful retrieval of motion data incorporating the time-space trade-off. We have proposed a system which is capable of dealing with large set of motion data employing an efficient database structure with improved performance. We have analyzed two motion representation techniques to realize the effectiveness of the system. Performance evaluation is performed by synthesized 3D human motions observed from eight camera directions. Finally, our results show that the proposed recognition scheme performs well for the captured motions. ©2010 IEEE.

DOI： 10.1109/ICARCV.2010.5707401

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79952382487&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper describes a novel technique for detecting a human and its body direction using HOG feature. The HOG feature is a well-known feature for the judgment of a human. But normally it contains the background feature, which gives negative influence on the judgment. This paper proposes the employment of the HOG feature based on a human model. The feature is also employed for detecting human body direction. Experimental results show effectiveness of the proposed technique compared to the conventional one. © 2010 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78649252315&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes an automatic human posture acquisition technique by a robot. Decrease of working population will be a vital problem in our society. Escaping from the work in danger is also important for a human being to raise QOL. To solve all theses issues, it may be quite useful to realize a robot which works in place of a human. Based on this background, our research aims at making a technique for a robot to acquire various human postures/motions automatically. In this paper, a technique is presented for obtaining human postures automatically by employing a human skeleton model and a stereo camera system. Performance of the technique is shown experimentally. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751531236&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Nowadays many persons are needed to observe images from surveillance cameras, because many surveillance cameras are installed in a town or in buildings. They are working under strain, because they must always watch the images from surveillance cameras to find a person with abnormal motion. Therefore reduction of the load is necessary. Many existing researches on human motion recognition are only recognition and the information of the result is not used. We assume that human motion is a set of basic motions. So, an overall motion can be understood using the recognition result of basic motions. The goal of the present research is to develop a method of human motion representation and translation using directional motion history images (DMHIs). In this paper, we describe a method of recognizing basic motions using the DMHIs. We perform the recognition by the Histograms of Oriented Gradients (HOG) feature. In the experiment, the number of bins and local area (cell) sizes for calculating the HOG feature are changed and the most suitable values are inspected. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751548634&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

We focus on human activity analysis so that an intelligent system (e.g., a robot) can easily understand some important activities and help thereby. Hence we present an improved method for activity analysis, called Medianfiltered Temporal Motion Segmentation (MfTMS) method, which can segment and understand motion temporally from the video sequence. It is based on the computation of optical flow and thereafter split it into four different channels. Later median filtering is applied and we produce four motion-history templates based on the directional motion vectors. Based on the total pixel volumes on these history templates and their related variations, various directions of the action primitives are segmented temporally. We conduct experimentations both indoor and outdoor environments and achieved sound performance. This segmentation method can assist an intelligent system or a robot to understand activities and take decisions afterwards. It is a simple but robust approach. © 2010 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866648695&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, arterio sclerosis obliterans (ASO) caused by the chronic ischemia of limbs increase remarkably. As one of the diagnosis methods of this disease, the image diagnosis such as MR image is used on medical fields. In this paper, we propose a method to develop medical diagnostic support system using fresh blood imaging (FBI) method, in which the structure of blood vessel regions in legs could be analyzed from non-enhanced MR images. As the method, an initial point of the vessel regions in the artery areas is acquired from weighted MIP images automatically. Then extend the artery areas using region growing method. Especially, we developed for the position relationship of the crossed blood vessel and the segment accuracy in peripheral artery areas that is improved by using weighted MIP images that the result could not be gotten in conventional MIP images. To evaluate our proposed method, we performed our proposed method and satisfactory results are achieved. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751514149&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper describes a novel technique for detecting a human body direction using SVM constructed by HOG feature selected by AdaBoost. HOG feature is well-known feature for the robust judgment of a human. We employ the feature for detecting a human body direction. We compared some feature selecting methods with the previous one. Experimental results show effectiveness of the proposed method. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751524891&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

We propose a technique for detecting pedestrians by employing stereo camera images and based on probabilistic voting. From a disparity map, each pixel on the image is voted on a depth map employing a 2-D Gaussian distribution. The region having the peak value in the vote is chosen as the foot of an object. The object is specified by a rectangle on the right image, which is referred to as the region of interest (ROI). This ROI is described by HOG features, and is judged by SVM if it contains a person. With an ROI containing a person, a Kalman filter is applied to track the person through successive image frames. The performance of the detection of people was evaluated by employing ground truth data. The ratio of people detected to the ground truth data, called the recall rate, was 80%. This is a satisfactory result. © 2010 International Symposium on Artificial Life and Robotics (ISAROB).

DOI： 10.1007/s10015-010-0843-2

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78651345756&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

we propose a technique for detecting pedestrians employing stereo camera images based on probabilistic voting. From a disparity map, each pixel on the image is voted on a depth map employing a 2-D Gaussian distribution. The region having a peak value of the vote is chosen as the foot of an object. The object is specified by a rectangle on the right image, which is referred to as a region of interest (ROI). This ROI is described by HOG features and it is judged by SVM if it contains a person. With the ROI containing a person, Kalman filter is applied to track the person through successive image frames. Performance of the detection of persons was evaluated employing a ground truth data. The rate of detected persons to the ground truth data, called a recall rate, was 80 %. This is a satisfactory result. © 2010 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866676321&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Tracking of multiple objects is more challenging than tracking a single object. Some problems arise in multiple object tracking that do not exist in single object tracking such as objects occlusion, appearing new object and disappearing already existed object, updating the occluded object, etc. In this paper, we present an approach to handle multiple objects tracking in the presence of occlusions, background clutter and appearance change. The occlusion is handled considering the predicted trajectories of the objects based on a dynamic model and likelihood measures. We propose also a target-model update conditions, ensuring the proper tracking multiple objects. The proposed method is implemented in the probabilistic framework such as particle filter in conjunction with a color feature. The particle filter has proven very successful for non-linear and non-Gaussian estimation problems. It approximates a posterior probability density of the state such as the object position by using samples or particles, where each state is denoted as the hypothetical state of the tracked object and its weight. The observation likelihood of the objects is modeled based on color histogram. The sample weight is measured based on Bhattacharya coefficient that measures the similarity between each sample's histogram with a specified target model. The algorithm can track the multiple objects in the present of occlusion and noise successfully. Some experimental results show the effectiveness of our method to track the multiple objects. © 2010 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866680510&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In this paper, we propose a CAD (Computer aided diagnosis) system to analyze the RA (rheumatoid arthritis) and osteoporosis by using image processing techniques from the CR images. To analyze the RA, we develop a segmentation method for phalanges in CR Image by use of DCT (Discrete Cosine Transform) for detection of temporal change. The temporal change is detected using the difference image between previous image and current one. The DCT is performed to emphasize the edge of the difference image. Finally, the phalanges are extracted by performing Snakes. The primary objective of this study is to segment phalanges by making temporal subtraction images. We apply our proposed technique to eight cases of CR images and satisfactory segmentation results are achieved. A new index that diagnoses the progress level of the disease of phalanges can be offered as a second opinion. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751512171&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Peripheral arterial disease (PAD) is characterized by lower extremities arterial obstruction due to athe rosclerosis and manifests in lower extremities as intermittent claudicating, limb ischemia, or gangrene. The dia gnosis of the PAD using magnetic resonance imaging (MRI) equipment without contrast medium is available as a useful visual screening in clinical practice. In this paper, we propose a new method for segmentation of arterial images, which are obtained from non-contrast enhanced magnetic resonance angiography (MRA) based on dot enhancement filter and 3-D region growing methods, and satisfactory experimental results are obtained. © 2010 ISAROB.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866686962&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Recently, multi detector row computed tomography (MDCT) has been introduced into medical fields. By the development of MDCT, images with high quality are provided into medical fields. So many related image processing techniques are proposed into medical image processing fields for extraction of abnormal area. In the medical image processing field, segmentation is one of the most important problems for analyzing the abnormalities and recognition of internal structures before the operation. For this reason, many approaches are proposed for detection of abnormal area on CT images. Before detection of abnormal areas, segmentation of organs in CT images is one of the most important problems for analyzing of disease. However, poor contrast, image noises and motion artifacts make this segmentation problem difficult in particular in cardiac region. Moreover, there are still no fully automatic segmentation methods for cardiac region on CT images. In this paper, we present automatic extraction technique for detection of cardiac region. Our proposed technique combines active shape model (ASM) and genetic algorithm (GA). We apply our proposed technique to five real CT images and satisfactory segmentation results are achieved. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751513157&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

CAD (Computer Aided Diagnosis) system that has 3-D viewer function to display the DICOM (Digital Imaging and Communication in Medicine) image has been developed in recent years. Now, the function of CAD system for artificial implant to which development is advanced includes the removal of artifact in the input DICOM image, region extraction of bone, major vessel, and nerve and mounting of function to presume place where implant is stricken. However, it is still required a software with high performance because of taking the processing time to do a series of operation before simulating by the strike of the implant on the operation on a clinical site. In this paper, we propose the function to extract the area of bone separately for teeth and mandibles automatically as addition to the system that can do a series of operation before simulating by the strike of the implant. In addition, we proposed also the function to extract each tooth separately. The main aim of this paper is to develop the integrative system for the dental screening that can be operated in a personal computer. ©ICROS.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78751559566&origin=inward

記述言語：日本語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

主要論文集（会議）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Korea   Seoul   2010年10月27日  -  2010年10月30日

主要論文集（会議）

記述言語：英語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：英語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taipei   2010年08月18日  -  2010年08月21日

主要論文集（会議）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taipei   2010年08月18日  -  2010年08月21日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taipei   2010年08月18日  -  2010年08月21日

主要論文集（会議）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taipei   2010年08月18日  -  2010年08月21日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Taiwan   Taipei   2010年08月18日  -  2010年08月21日

主要論文集（会議）

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

China   Chengdu   2010年07月07日  -  2010年07月10日

DOI： 10.1109/ICCSIT.2010.5563884

Scopus

記述言語：英語   掲載種別：研究論文（その他学術会議資料等）

主要雑誌

DOI： 10.3844/jcssp.2010.920.927

Scopus

記述言語：英語   掲載種別：研究論文（学術雑誌）

主要雑誌

記述言語：日本語   掲載種別：研究論文（学術雑誌）

記述言語：日本語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：日本語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：英語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：英語   掲載種別：研究論文（学術雑誌）

主要雑誌

DOI： 10.1007/s10015-010-0841-4

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

America   California ,Santa Barbara   2010年03月20日  -  2010年03月25日

DOI： 10.1109/FG.2011.5771342

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

America   San Francisco   2010年03月13日  -  2010年03月18日

DOI： 10.1109/CVPRW.2010.5543160

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Alabama   Auburn   2010年03月14日  -  2010年03月16日

DOI： 10.1109/ICIT.2011.5754412

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

名古屋   2010年03月08日  -  2010年03月11日

記述言語：英語   掲載種別：研究論文（学術雑誌）

Various recognition methodologies address to recognize and understand varieties of human activities. However, motion self-occlusion due to motion overlapping in the same region is a daunting task to solve. Various motion-recognition methods either bypass this problem or solve this problem in complex manner. Appearance-based template matching paradigms are simpler and hence these approaches faster for activity analysis. In this paper, we concentrate on motion self-occlusion problem due to motion overlapping in various complex activities for recognition. In the Motion History Image (MHI) method, the self-occlusion is evident and it should be solved. Therefore, this paper compares our directional motion history image concept with basic the Motion History Image, Multi-level Motion History representation and Hierarchical Motion History Histogram representation to solve the self-occlusion problem of basic the Motion History Image representation. We employ some complex aerobics and find the robustness of our method compared to other methods for this self-occlusion problem. We employ seven higher order Hu moments to compute the feature vector for each activity. Afterwards, k-nearest neighbor method is utilized for classification with leave-one-out paradigm. The comparative results clearly demonstrate the superiority of our method than other recent approaches. We also present several experiments to demonstrate the performance and strength of the DMHI method in recognizing various complex actions. © 2010 ACADEMY PUBLISHER.

DOI： 10.4304/jmm.5.1

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78651574909&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年02月04日  -  2010年02月06日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年02月04日  -  2010年02月06日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年02月04日  -  2010年02月06日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年02月04日  -  2010年02月06日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年02月04日  -  2010年02月06日

記述言語：英語   掲載種別：研究論文（学術雑誌）

主要雑誌 代表的研究業績

記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.1007/s10015-010-0762-2

Scopus

記述言語：英語   掲載種別：研究論文（学術雑誌）

DOI： 10.1007/s10015-010-0764-0

Scopus

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年01月27日  -  2010年01月29日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年01月27日  -  2010年01月29日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

別府   2010年01月27日  -  2010年01月29日

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Indonesia   Jakarta   2010年01月06日  -  2010年01月08日

主要論文集（会議） 代表的研究業績

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes a novel 3-D modeling technique of an object. The existent 3-D modeling techniques such as stereo vision or factorization recovers 3-D shapes of part of an object observed commonly from multiple orientations. This is followed by registration among recovered partial shapes in order to obtain an entire 3-D model of the object, resulting in some geometrical errors in the linked shape. The proposed technique creates entire shape of an object simultaneously without partial shape registration by the employment of virtual see-through cameras that surround the object and observes its rear part as well as the frontal part. All the surrounding cameras are calibrated initially using the captured images. The frontal side of an object concerned recovers its 3-D shape by the factorization method applied to the image data obtained from the frontal cameras. The recovered frontal shape is virtually projected onto the image planes of the rear cameras. This procedure realizes that all the surrounding cameras observe all the shape of the object, irrespective of frontal or rear. Then the factorization method is applied again to this situation to yield an entire 3-D model of the object. A nonlinear optimization method is employed in the technique to achieve the modeling of higher precision. The proposed technique is applied to the 3-D modeling of a mini basket ball match and their activity is successfully modeled in a 3-D way. © 2010 SICE.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78649296463&origin=inward