研究者詳細 - 吉田　香

論文 - 吉田　香

分割表示114 件中 1 - 20 件目／全件表示 >>

The Impact of Large Sample Datasets on Hand Gesture Recognition by Hand Landmark Classification 査読有り国際誌

AHMAD Khawaritzmi Abdallah, SILPANI Dian Christy, YOSHIDA Kaori

International Journal of Affective Engineering ( 日本感性工学会 ) 22 ( 3 ) 253 - 259 2023年01月

担当区分：最終著者記述言語：英語掲載種別：研究論文（学術雑誌）

Today, almost all aspects of life are influenced by technology to simplify these aspects of life. Researching how humans and computers interact is very important to build a good model for better human-computer interaction in the future. Using technology such as computer vision, we can now collect the information we are looking for directly from a human. Humans can use many kinds of modalities to interact with computers. Hands are perhaps the largest source of body language information after the face. To understand the gesture’s meaning, we can use MediaPipe Hands, developed at Google LLC, as a method to track and recognize human hands. However, if we want to understand some kinds of hand gestures using MediaPipe Hands, we need to create a condition using if-else manners. This research tried to collect the wide variety of each hand gesture using the 21 key points in x, y, and z coordinates as a feature. We chose Support Vector Machine (SVM) and Artificial Neural Network (ANN) as classifiers to validate the impact of large sample datasets. This research found that ANN is the best among all the methods we used as a classifier method for the 3D value of 21 key points from the hand skeleton. The accuracy and F1-score from ANN were 98.4% accuracy and 98.2% F1-Score, respectively, representing the best performance for each class of all the methods we used in this research.

DOI： 10.5057/ijae.ijae-d-22-00022

CiNii Research

A Feasibility Study on Hand Gesture Intention Interpretation Based on Gesture Detection and Speech Recognition 査読有り国際誌

Silpani Dian Christy, Suematsu Keishi, Yoshida Kaori

Journal of Advanced Computational Intelligence and Intelligent Informatics ( 富士技術出版株式会社 ) 26 ( 3 ) 375 - 381 2022年05月

担当区分：最終著者記述言語：英語掲載種別：研究論文（学術雑誌）

In this paper, considering the characteristics of hand gestures that occur during a natural, unscripted human conversation, we report the results of experiments involving the data analysis of the video recordings of experiments. The recorded conversations between two subjects and the assistant in the experiments (the first author) are analyzed for findings that might be useful to future human-robot interaction (HRI), findings regarding the recognition of hand gestures in natural conversations. In the experiment, hand gestures that appeared naturally during the conversations were manually selected, and their elapsed times and number of appearances were analyzed. In addition, these unscripted conversations were analyzed with a voice recognition system, and the meanings of the hand gestures that appear in the recorded video were interpreted. It was observed that some gestures were made consciously and others unconsciously. Some gestures appeared multiple times with similar meanings and the same intention, especially gestures involving pointing at something to emphasize a sentence or word. Other observations included that there were times when (i) each subject made different hand gestures to convey the same meaning, (ii) both subjects made the same hand gesture but intended different meanings, and (iii) a specific gesture was made by the same subject but with different intended meanings. After considering the results of these experiments, we conclude that it is better to combine gestures with the utterances at the time in order to interpret their intended meanings correctly.

DOI： 10.20965/jaciii.2022.p0375

Scopus

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85132786254&origin=inward

Hand Gesture Recognition by Hand Landmark Classification 査読有り国際誌

AHMAD Khawaritzmi Abdallah, SILPANI Dian Christy, YOSHIDA Kaori

International Symposium on Affective Science and Engineering ( 日本感性工学会 ) ISASE2022 ( 0 ) 1 - 4 2022年01月

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

These years, computer vision technology has been rapidly advanced. On the other hand, customer satisfaction is essential for the industry to improve its service but use traditional methods ineffectively. Using technology such as computer vision, we can now collect the information we are looking for directly from a human. Humans can use many kinds of modalities to interact with computers. Hands are perhaps the largest source of body language information after the face. To understand the gesture's meaning, we can use MediaPipe Hands developed at Google LLC as a method to track and recognize human hands. However, if we want to understand some kinds of hand gestures using MediaPipe Hands, we need to create a condition ourselves by using if-else conditions. This research tried to collect the many varieties of each hand gesture using the 21 key points in x, y, and z coordinates as a feature. As classifiers, we chose the support vector machine (SVM) and the artificial neural network (ANN). This research found that SVM using a polynomial kernel is the best among all of the methods we used as a classifier method for the 3D value of 21 key points from the hand skeleton. The accuracy and F1-score from SVM using a polynomial kernel were 86.26% accuracy and 82% F1-score, respectively, representing the best performance for each class of all the methods we used in this research.

DOI： 10.5057/isase.2022-c000026

CiNii Research

特集：「地域における AI 社会実装」特集「地域における AI 社会実装」にあたって

吉田香, 諏訪博彦, 清田陽司

人工知能 ( 一般社団法人人工知能学会 ) 37 ( 2 ) 170 - 171 2022年01月

Calibration Cost Reduction of Indoor Localization Using Bluetooth Low Energy Beacon 査読有り国際誌

As M., Shimizu H., Benaissa B., Yoshida K., Koppen M.

Journal of Advanced Computational Intelligence and Intelligent Informatics ( 富士技術出版株式会社 ) 26 ( 1 ) 97 - 106 2022年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Indoor localization based on Bluetooth low energy (BLE) beacons has been rapidly developed, and many approaches have been developed to achieve higher estimation accuracy. In these methods, the received signal strength (RSS) is the input. However, the measurement of indoor environments is affected easily; the signal may be reflected and attenuated by obstacles such as the human body, walls, and furniture, which creates a challenge for methods based on signal mapping. In this study, BLE signal characteristics are investigated in an indoor localization setting. An experiment is performed using one BLE beacon and multiple receivers installed at different wall and ceiling positions. The raw RSS is observed, and the relationship between the BLE beacon signal strength characteristics against the human body effect as well as the receiver's placement in the observation area are discussed. Signal mapping is performed, where the signal strength is measured from all receivers simultaneously. The position estimation accuracy is examined based on different data scenarios. The results show that the estimation position estimated by the BLE beacon based on extensive BLE beacon data does not affect the estimation accuracy.

DOI： 10.20965/jaciii.2022.p0097

Scopus

CiNii Article

CiNii Research

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124460770&origin=inward

A First Prototype of an Emotional Smart Speaker 査読有り国際誌

Enrique A. de la Cal, Alberto Gallucci, José Ramón Villar, Kaori Yoshida, Mario Koeppen

Advances in Intelligent Systems and Computing 1401 304 - 313 2021年09月

A Feasibility Study on Hand Gesture Recognition in Natural Conversation 査読有り

Silpani D.C., Suematsu K., Yoshida K.

2021 5th IEEE International Conference on Cybernetics, CYBCONF 2021 85 - 90 2021年06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）

This paper proposes to approach natural hand gestures that occur in natural conversations. Hand gesture recognition in Human-Robot Interaction is considered a preliminary study for the next step, based on the vision system of a humanoid robot. We experimented as a feasibility study on hand gestures recognition in natural conversation. The conversations between an experimental assistant (the first author) and two different subjects (both female) were recorded in the experiment. Each conversation includes two people, one of whom is the subject of the hand gesture detection that appears during the conversation. We manually captured the hand gestures that appear naturally during the session and analyzed their duration and frequency of occurrence from each video. As a result of conversations with two subjects, it was clear that hand gestures significantly impact people's conversations. We observed that hand gestures with the same word or similar meaning appeared several times with the same intention, such as pointing at something and emphasizing a sentence or word. The topic of natural hand-gesture recognition by a robot will be a challenge. Cultural diversity makes the scope of natural hand-gesture broad, but this can be considered a topic in developing the knowledge of human-robot interaction in the future.

DOI： 10.1109/CYBCONF51991.2021.9464141

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85113788192&origin=inward

より良い未来のための社会実装

吉田香

人工知能 ( 一般社団法人人工知能学会 ) 36 ( 3 ) 335 - 335 2021年01月

Smart watering system based on framework of low-bandwidth distributed applications (LBDA) in cloud computing 査読有り

Sirimorok N., As M., Yoshida K., Köppen M.

Advances in Intelligent Systems and Computing 1263 AISC 447 - 459 2021年01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）

The estimation of soil moisture content is required for agriculture, mainly to build the irrigation scheduling model. In this study, we present a smart watering system to deal with various factors derived from the stochastic information in the agricultural operational, i.e., air temperature, air humidity, soil moisture, soil temperature, and light intensity. The methodology consists of exploiting the Internet of Things (IoT) using Low-Bandwidth Distributed Applications (LBDA) in cloud computing to integrate the real data sets collected by the several sensor technologies. We conducted experiments for the watering system used two types of soil and different plant. Here, the Long Short Term Memory Networks (LSTMs) approach in deep learning techniques used to build smart decisions concerning watering requirements and deal with heterogeneous information coming from agricultural environments. Results show that our models can effectively improve the prediction accuracy for the watering system over various soil and plants.

DOI： 10.1007/978-3-030-57796-4_43

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85090100059&origin=inward

Simple Meta-optimization of the Feature MFCC for Public Emotional Datasets Classification 査読有り国際誌

de la Cal E., Gallucci A., Villar J.R., Yoshida K., Koeppen M.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12886 LNAI 659 - 670 2021年01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）

A Speech Emotion Recognition (SER) system can be defined as a collection of methodologies that process and classify speech signals to detect emotions embedded in them [2]. Among the most critical issues to consider in an SER system are: i) definition of the kind of emotions to classify, ii) look for suitable datasets, iii) selection of the proper input features and iv) optimisation of the convenient features. This work will consider four of the well-known dataset in the literature: EmoDB, TESS, SAVEE and RAVDSS. Thus, this study focuses on designing a low-power SER algorithm based on combining one prosodic feature with six spectral features to capture the rhythm and frequency. The proposal compares eleven low-power Classical classification Machine Learning techniques (CML), where the main novelty is optimising the two main parameters of the MFCC spectral feature through the meta-heuristic technique SA: the n_mfcc and the hop_length. The resulting algorithm could be deployed on low-cost embedded systems with limited computational power like a smart speaker. In addition, the proposed SER algorithm will be validated for four well-known SER datasets. The obtained models for the eleven CML techniques with the optimised MFCC features outperforms clearly (more than a 10%) the baseline models obtained with the not-optimised MFCC for the studied datasets.

DOI： 10.1007/978-3-030-86271-8_55

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115856860&origin=inward

Hand Position Tracking Correction of Leap Motion Controller Attached to the Virtual Reality Headset 査読有り

BELANOVÁ Dorota, YOSHIDA Kaori

International Journal of Biomedical Soft Computing and Human Sciences: the official journal of the Biomedical Fuzzy Systems Association ( バイオメディカル・ファジィ・システム学会 ) 25 ( 1 ) 29 - 37 2020年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Considering the increasing number of patients with disabilities, especially of upper extremity, we decided to develop hand rehabilitation virtual reality system for post-stroke patients using leap motion controller. However, as the device has not met the required accuracy of hand position tracking, we created a real-time low-computational cost correction algorithm lessening the problem. This paper outlines the challenges of hand position tracking accuracy identifying the parts contributing to the overall error, proposes a method of correction suitable for each of the parts separately and brings the parts together into an algorithm maximizing the hand position correction. The algorithm was tested on twelve healthy participants in two series of four tasks. Results showed our algorithm improved the hand position accuracy considerably in all the cases.

DOI： 10.24466/ijbschs.25.1_29

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130007998938

The Concept of Intelligent Space with a Robot 査読有り

BELANOVÁ Dorota, MACH Marián, SINČÁK Peter, YOSHIDA Kaori

International Symposium on Affective Science and Engineering 2019 ( 0 ) 1 - 4 2019年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Intelligent space is where people may easily interact with computers and robots, changing the environmental conditions based on users' needs to create more comfortable and satisfying surroundings. In our proposed system, we would integrate an intelligent space with a robot that would act as an active part of the environment. The system is illustrated in a few caregiver scenarios, showing the actions of a robot. The skills of a robot are explained, presenting the advantages and contributions of the system.

DOI： 10.5057/isase.2019-C000024

Kyutacar

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130007656892

Tensor self-organizing map for kansei analysis 査読有り

Itonaga K., Yoshida K., Furukawa T.

Proceedings - 2018 Joint 10th International Conference on Soft Computing and Intelligent Systems and 19th International Symposium on Advanced Intelligent Systems, SCIS-ISIS 2018 796 - 801 2018年12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）

© 2018 IEEE. In Kansei analysis, impressions of various objects are commonly measured using evaluation words. When using this approach, it is necessary to examine all combinations of three elements: subjects, objects, and evaluation words. However, the exhaustive analysis required is not an easy task because of the enormous number of combinations. Additionally, if it is necessary to reveal the relationship between the impressions and physical features of objects such as colors or shapes, the number of combinations increases enormously and the task becomes unrealistic. In this paper, we introduce a method called the tensor self-organizing map (TSOM) that visualizes the relationships between the elements. We applied the TSOM to Kansei analysis of landscape images and studied how the impressions were dependent on the subjects. We also investigated the relationships between these subject-dependent impressions and the physical features. Through these experiments, we demonstrate that the TSOM can be a useful tool for Kansei analysis.

DOI： 10.1109/SCIS-ISIS.2018.00132

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85067131361&origin=inward

An Efficient Replay Method to Prevent SURF-Based Face Detection 査読有り

Dadkhah S., Koppen M., Sadeghi S., Yoshida K.

Proceedings - 2018 5th Asia-Pacific World Congress on Computer Science and Engineering, APWC on CSE 2018 18 - 25 2018年12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）

© 2018 IEEE. During the past couple of years, a considerable number of digital images are gathered by search engines, or uploaded by users in different social websites. These images could contain a variety of variables such as different faces, scenes, and unusual objects. The advancement of digital face recognition algorithms is sensational, recognition algorithms used by some of the social websites are even stronger than most secret service establishments. However, there is always a concern about how to protect the user's privacy when they do not want to be recognized automatically by these services. In this research, a SURF-based face recognition algorithm along with a featurebased attack to mislead this type of scheme is proposed. The recognition algorithm use different machine learning technique along with second layer descriptor counting function which makes it more powerful. However the primary focus of this research is to provide a method which can mislead this type of recognition systems with a smallest possible degradation to the image representation. The experimental results show that the proposed plan is successful to fool the SURF algorithm and its recognition system by 70%.

DOI： 10.1109/APWConCSE.2018.00012

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85074295713&origin=inward

Updatable indoor localization based on BLE signal fingerprint 査読有り

Benaissa B., Yoshida K., Koppen M., Hendrichovsky F.

Proceedings of the 2018 International Conference on Applied Smart Systems, ICASS 2018 2018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）

© 2018 IEEE. The need for a robust indoor localization solution is mounting in the recent few year, as the IoT technology is growing fast. In this paper, we discuss existing indoor localization approaches, focusing on the signal strength fingerprint approaches, to which belongs the presented study and based on Bluetooth Low Energy signal. The proposed approach employs the Radial Basis Functions (RBF) to create model that describe the real signal strength-position relationship. The present method is fully implemented on a smart phone, namely, offline data collection, model computation and online position estimation. This allow updating the model by collecting more signal data, therefore improving the localization accuracy with small efforts. The presented results shows that updating the model effectively improves the accuracy of estimation.

DOI： 10.1109/ICASS.2018.8652035

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85063516885&origin=inward

上半身動作を用いたロボットの回避方向意図の伝達査読有り

石川友紀也, 末松圭史, 大坪正和, 吉田香

日本知能情報ファジィ学会論文誌 30 ( 5 ) 709 - 716 2018年10月

Path planning on robot based on D∗ lite algorithm 査読有り

Belanova D., Mach M., Sinčák P., Yoshida K.

DISA 2018 - IEEE World Symposium on Digital Intelligence for Systems and Machines, Proceedings 125 - 130 2018年10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）

© 2018 IEEE. The increasing need of autonomous behavior of robots in fields of science and technology formed the requirement for path planning implemented by the robot without the human assistance. In this paper, D∗ Lite, which is a path planning graph-based algorithm, was used in order to compute the shortest path from a start to goal point in a real environment and make a Pepper robot move in a computed trajectory. The movement of robot was conducted in a static environment, with the map of the environment already known. This paper is a first step in the research focusing on a creation of a so-called intelligent workspace.

DOI： 10.1109/DISA.2018.8490605

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85056782104&origin=inward

An efficient ward-based copy-move forgery detection method for digital image forensic 査読有り

Dadkhah S., Koppen M., Sadeghi S., Yoshida K., Jalab H., Manaf A.

International Conference Image and Vision Computing New Zealand 2017-December 1 - 6 2018年07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）

© 2017 IEEE. In recent years, with the improvement of digital manipulation tools, image forgeries have become a genuine social issue. Copy-move forgery is an important type of digital manipulation which is done by duplicating some parts of the image with the purpose of hiding specific information. Verifying the authenticity of the digital image can become vital for different forensic organizations. In this paper, an efficient copy-move detection algorithm with three level of Ward linkage based clustering is proposed. The proposed algorithm extracts the local image features with the help of Scale-invariant feature transform algorithm (SIFT). In the proposed method, a Ward-based clustering scheme followed by three levels of false-descriptor elimination is applied. The proposed elimination system utilizes Euclidean distance to improve the accuracy of the region detection scheme. The efficiency of the proposed algorithm is examined against several different post-processing attacks such as multilevel forgeries, noises, and JPEG compression. Experimental results illustrate that the proposed algorithm improves the detection accuracy by eliminating false descriptors.

DOI： 10.1109/IVCNZ.2017.8402472

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85050008851&origin=inward

Bad Ai: Investigating the Effect of Half-Toning Techniques on Unwanted Face Detection Systems 査読有り

Dadkhah S., Köeppen M., Sadeghi S., Yoshida K.

2018 9th IFIP International Conference on New Technologies, Mobility and Security, NTMS 2018 - Proceedings 2018-January 1 - 5 2018年03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）

© 2018 IEEE. Currently, automatic face detection systems are widely used in various social media. The improvement of the automated intelligent face recognition can easily result in the invasion of the user 's privacy if the predefined functions for their so- called artificial intelligent systems are designed with no respect to the user privacy, this means that users do not get the right to prevent these types of automatic face detection. The problem does not stop here, and recently some organization attempted to use automated face recognition system as a digital personal judgment tools. For instance, judging if a person is criminal only based on the attributes of his/her face. In the present technology, spreading the personal information such as individual digital images are fast and inevitable. Thus, in this paper, some of the techniques to evade automatic face recognition is investigated. The focused of this article is to highlight the advantages of different Half-Toning algorithms concerning avoiding unwanted automated face detection/recognition. In the experimental phase, the result of varying filtering and modification algorithms compared to the proposed filtering technique.

DOI： 10.1109/NTMS.2018.8328726

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85050981631&origin=inward

Phone Application for Indoor Localization Based on Ble Signal Fingerprint 査読有り

Benaissa B., Hendrichovsky F., Yishida K., Köppen M., Sinčák P.

2018 9th IFIP International Conference on New Technologies, Mobility and Security, NTMS 2018 - Proceedings 2018-January 1 - 5 2018年03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）

© 2018 IEEE. Indoor position of devices and persons is a very valuable information for context-aware devices and human activity recognition. The need of such information grows bigger by the advancement of IoT technologies. In this paper, we discuss existing indoor positioning approaches and their applications, give specific interest to the signal strength fingerprint approaches, to which belongs the presented paper. Bluetooth Low Energy Beacons are used as the source of the signal, and the Radial Basis Functions method is employed to create a model that relates to signal strength and position. The proposed approach allows the possibility of making all the computation on a smartphone application, namely, offline data collection and model computation, and online position estimation. Besides, it allows updating the model later.

DOI： 10.1109/NTMS.2018.8328729

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85047863726&origin=inward

　前のページ- 次のページ　

<吉田 香>

論文 - 吉田 香

<吉田　香>

論文 - 吉田　香