研究者詳細 - 神谷　亨

論文 - 神谷　亨

分割表示757 件中 181 - 200 件目／全件表示 >>

Image Registration Method for Temporal Subtraction Based on Salient Region Features 査読有り

Sato S., Lu H., Kim H., Murakami S., Ueno M., Terasawa T., Aoki T.

EAI/Springer Innovations in Communication and Computing 13 - 20 2021年01月

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）

While diagnosis imaging is an indispensable technology in medical field, it causes an increase in the burden to radiologists. In recent years, computer-aided diagnosis (CAD) system for supporting a radiologist has been developed to solve this problem. Temporal subtraction, which is one of CAD, is a technique to generate images emphasizing temporal changes in lesions and facilitates diagnosis of radiologists. To make a temporal subtraction image, image registration technique is required. In this chapter, we propose an image registration method for image registration of current image and previous image to generate temporal subtraction images in a short time. The proposed method consists of three steps: (i) segmentation of the region of interest (ROI) using position information of the spine based on anatomical information, (ii) using global image matching to select pairs of previous image and current image in which the same portion is depicted, and iii) final image matching based on salient region features (SRF). We performed our registration technique to the synthetic data and confirmed usefulness of the proposed method. The rotated synthesis image gives TP 100.0% and FP 12.16%. The synthesis image obtained by applying a Gaussian filter gives TP 70.40% and FP 0.00%. The synthesis image obtained by adding artificial pseudo-lesion region gives TP 99.45% and FP 17.89%. The synthesis image obtained by adding random noise of 5% gives TP 83.05% and FP 16.95%. Furthermore, radiologist conducted comparative experiments without and with temporal subtraction images created by proposed method. As a result, radiologist showed high reading performance by using temporal subtraction images.

DOI： 10.1007/978-3-030-46032-7_2

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091327306&origin=inward

Extreme ROS Reality: A Representation Framework for Robots Using Image Dehazing and VR 査読有り

Ueda A., Lu H., Kim H.

EAI/Springer Innovations in Communication and Computing 21 - 32 2021年01月

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）

It is necessary to work in response to various situations when using robots at disaster areas. So, the complete automation remote control technique of the robot is essential. In recent years, the remote control technique of the robot is mainly focused on recognizing the working environment of the robot from the 2D monitor and performing the operation. In order to operate the robot correctly, a corresponding technique and sufficiently trained engineers are required. To this end, a new remote control technology that can operate more easily is required. Also, depending on the disaster environment that working at night or indoors is assumed, it becomes difficult to recognize the work environment correctly from the image by the sensor such as the camera due to insufficient light source. In this chapter, in order to solve these problems, as a development of new remote control technology, we propose the method of reproducing the working environment and the image improvement method under the low light source environment. In this chapter, we reproduce the working environment of remote areas by using virtual reality. In addition, we improve the image under low light source environment using haze removal algorithm. The work environment of the robot is acquired by RGBD sensor. Image improvement processing is performed on the obtained color image, and the image data is transmitted to a remote place where the operator is located via WebSocket. The experiments show that our method reproduces the working environment of the robot as point group to the virtual reality space and examine its usefulness.

DOI： 10.1007/978-3-030-46032-7_3

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091329616&origin=inward

An Adaptable Feature Synthesis for Camouflage 査読有り

Li G., Yu Z., Chang J., Kim H.

EAI/Springer Innovations in Communication and Computing 81 - 91 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）

Military camouflage should be various in order to quickly adapt environmental changes of battlefield. The elusiveness is depressed if the constant camouflage patterns of the clothes and vehicles are used, as applying in the most of current military. In this paper, we propose a texture synthesis method from an image using convolutional neural networks, which has an effect on the camouflage generation on a 3D surface. We use the latest advances in style transitions in 2D images and map characteristic textures onto 3D surfaces according to geometric feature. This allows us to make an adaptive texture map of the 3D objects, even in the case of complex topologies.

DOI： 10.1007/978-3-030-46032-7_8

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091331451&origin=inward

A Classification Method for Magnetic Particle Testing Image Using U-Net 査読有り

Moritsuka S., Kamiya T.

International Conference on Control, Automation and Systems 2021-October 2047 - 2050 2021年01月

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）

Magnetic Particle Testing (MPT) is a method for determining the presence or absence of a defect by magnetizing the object to be inspected and sprinkling magnetic powder, which is absorbed by the defective part such as a crack and appears as a magnetic powder pattern, which is then evaluated by a specialist. By using the MTP, inspection can be performed without breaking the object to be inspected. However, there are some problems such as the possibility of overlooking defects. In this paper, to solve the problems we develop a classification method of defect images by deep learning for the automation of MPT. The proposed method is based on the structure of U-Net, which has excellent segmentation capability in image processing, and performs segmentation using an improved model that adds convolutional layers to U-Net. Then, an algorithm that combines the result with the last part of the encoder of U-Net is used to discriminate the presence or absence of defects. Using this method, defects were classified from the images obtained during MPT. The results showed that Accuracy of 85.8%, TPR of 65.2%, and FPR of 13.8%.

DOI： 10.23919/ICCAS52745.2021.9650052

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124232188&origin=inward

Underwater image super-resolution using SRCNN 査読有り国際誌

Ooyama S., Lu H., Kamiya T., Serikawa S.

Proceedings of SPIE - The International Society for Optical Engineering 11884 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）

In recent years, energy minerals have become more important due to the rapid industrialization worldwide. Due to the rapid industrialization on a global scale, there is a shortage of mineral resources, and there are more opportunities to rely on alternative energy sources. Therefore, the exploration of marine resources, which are abundant in the ocean, is being promoted. However, it is dangerous and impractical for humans to dive and search for marine resources by hand. Therefore, it is possible to proceed with underwater exploration safely by having a robot do the work instead. Robots have been used as a mainstream search tool in the underwater environment due to the existence of various hazardous environmental conditions. However, there are several problems associated with robot control in underwater environments, one of which is poor visibility in the water. One of the problems is the poor visibility in the water. To improve the visibility in the water, we are trying to increase the resolution of underwater images by using super-resolution technology. In this paper, we conduct experiments using SRCNN, which is a basic super-resolution technique for underwater images. In addition, we investigate the effectiveness of "Mish", which has been attracting attention in recent years for its potential to surpass the performance of "ReLU", although "ReLU"is a typical activation function of neural networks, on SRCNN.

DOI： 10.1117/12.2603761

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85120500372&origin=inward

Shape restoration by shadow information and photometric stereo 査読有り国際誌

Kameda S., Lu H., Kamiya T., Serikawa S.

Proceedings of SPIE - The International Society for Optical Engineering 11884 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）

Virtual Reality (VR) systems are become popular in recent years, and the capture 3D objects from the real world have been studied. 3D objects have large and complex data. In this paper, we propose a novel method that use the shadow information and Photometric stereo with their surface from one point of view to recover the 3D shapes. The experimental results show that the proposed method performs well accuracy.

DOI： 10.1117/12.2604193

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85120506573&origin=inward

Image quality improvement using local adaptive neighborhood-based dark channel prior 査読有り国際誌

Onoyama T., Lu H., Soomro A.A., Mokhtar A.A., Kamiya T., Serikawa S.

Proceedings of SPIE - The International Society for Optical Engineering 11884 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）

In-vehicle cameras and surveillance cameras are used in many situations in our daily lives. Visibility degradation in foggy environments is caused by the scattering of reflected light from real objects by minute water droplets or fog in the medium through which light passes. The degree of degradation depends on the density of suspended microparticles existing between the observed object and the observation point in the medium. In general, the farther the object is from the camera, the more it is affected by the fog. The purpose of image de-fogging is to improve the clarity of an object by removing the effects of fog in the image.

DOI： 10.1117/12.2603771

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85120496362&origin=inward

Incorporating Ghost Module into RCAN for Super-Resolution of Satellite Images 査読有り国際誌

Ikeda H., Li G., Kamiya T.

International Conference on Control, Automation and Systems 2021-October 2059 - 2063 2021年01月

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 explosion of amount of low cost satellites, satellite images have been widely used for many non-military applications, such as agriculture, landscape, and recognition of environment. Improving the image resolution to mine useful information becomes one of the immediate problems. Therefore, it is expected to improve the recognition accuracy by increasing the resolution of satellite images. Recently, deep learning technique has been proposed to increase the resolution of images. However it requires a large number of learning parameters, which results in huge computational cost. To overcome this problem, we develop a new deep learning model based on ghost module to reduce the parameters while maintaining the quality of results. We utilized Google Earth Pro satellite imagery for the network training and testing. Comparing to the classical convolutional neural network module based methods, the number of parameters used in our model was reduced 49.31 % but keeping the same level of Peak Signal-to-Noise Ratio (24.1578) and Structural Similarity (0.7174).

DOI： 10.23919/ICCAS52745.2021.9649982

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124232834&origin=inward

Extraction of Cervical Lymph Nodes Based on Three-Dimensional Image Registration 査読有り国際誌

Shime N., Kamiya T., Ishida T.

International Conference on Control, Automation and Systems 2021-October 1792 - 1795 2021年01月

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 difficulties of cancer is the metastasis of cancer through the lymph nodes. Therefore, early detection and treatment of cancerous is important. Visual screening is one of useful tool for diagnosing of the cancer. However, the number of images obtained at a time is large, and the burden on the reading physician is increasing. Furthermore, reading of image is based on the subjective judgment of the physician, which may lead to different diagnostic results. To solve these problems, computer aided diagnosis (CAD) systems have been attracting attention in recent years. One of the CAD systems is the temporal subtraction image technology. In this paper, we propose an image alignment method for generating temporal subtraction images from images taken before and after contrast agent was administered to the cervical lymph nodes of the same subject. In addition, we propose an image analysis method that suppresses overextraction of lymph node candidate regions on the temporal subtraction image based on the features of lymph nodes. We applied the proposed method to the CT images of three sets and compared the temporal subtraction images with the final lymph node extraction images, and confirmed that the proposed method can suppress the overextraction of lymph node regions.

DOI： 10.23919/ICCAS52745.2021.9649797

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124183732&origin=inward

Environment Recognition from A Spherical Camera Image Based on DeepLab v3+ 査読有り国際誌

Nishida Y., Li Y., Kamiya T.

International Conference on Control, Automation and Systems 2021-October 2043 - 2046 2021年01月

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）

The number of users of electric wheelchairs has been increasing in recent years because it is easy to operate the electric wheelchair and do not require physical strength. However, the traffic accidents are also increasing because of the large number of wheelchairs. The development of autonomous electric wheelchairs is expected to reduce the risk of accidents and improve the convenience of electric wheelchairs. Environmental recognition is essential for the development of autonomous electric wheelchairs. In this paper, we propose a method for recognizing roads, sidewalks, buildings, electric wheelchair drivers, poles, electric wheelchairs, vegetation, curbs, sky, pedestrians, lanes, cars, steps, and bicycles. For recognizing those objects, we use a panoramic image acquired from a spherical camera. As the machine techniques, we use DeepLab v3+, a semantic segmentation algorithm based on Convolutional Neural Network (CNN). In the proposed method, a new CNN model is constructed by adding deformable convolution, SE-block, and MobileNet v2 to DeepLab v3+ into the original DeepLab v3+. In the experiment, IoU 38.8% and Dice of 46.7% were obtained.

DOI： 10.23919/ICCAS52745.2021.9649846

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124192140&origin=inward

Determination of Abnormality of IGBT Images Using VGG16 査読有り国際誌

Ogawa T., Watanabe A., Omura I., Kamiya T.

International Conference on Control, Automation and Systems 2021-October 2055 - 2058 2021年01月

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 power device is a semiconductor device for power control used for power conversion such as converting direct current to alternating current and alternating current to direct current. It is widely used such as refrigerators, air conditioners which is implemented electronic components that are closely related to our daily lives. Therefore, high reliability and safety are required, and power cycle tests are conducted for the purpose of evaluating them. In the conventional test, there is a problem that it is difficult to perform analysis because sparks are generated during the test and the device is severely damaged after the test. To solve this problem, a new technology has been developed that adds ultrasonic that enable internal observation during the test. However, there are remains a problem that the method for analyzing the ultrasonic image obtained in the new technology has not been established. Also, few abnormal images are obtained in the test. In this paper, we propose a method for detection of abnormal devices based on CNN. Especially, we implement a Cycle-GAN to extend the abnormal data and classify the known image based on improved VGG16. As an experimental result, classification accuracy of Precision = 97.06%, Recall = 93.58%, F-measure = 95.17% were obtained.

DOI： 10.23919/ICCAS52745.2021.9650029

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124191152&origin=inward

Detection of the root resorption from panoramic X-ray images using deep metric learning 査読有り国際誌

Tamura K., Kamiya T., Oda M., Tanaka T., Morimoto Y.

International Conference on Control, Automation and Systems 2021-October 1800 - 1803 2021年01月

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）

Root resorption is a pathological process characterized by the loss of tooth roots because of inflammation induced by bacterial infection, trauma, physical or chemical irritation. As a result, the development of periodontal disease, increased susceptibility to infection and crooked teeth. In the worst case, it can lead to tooth extraction. Root resorption is often caused by pressure during orthodontic treatment. The presence of root resorption should be checked regularly during orthodontic treatment, as it often occurs. It is necessary to check for root resorption periodically during orthodontic treatment. However, it is difficult to detect the root resorption using a panoramic radiograph. As a result, root resorption is often latent and goes undetected. In this paper, we propose an image analysis method based on deep learning technique for detecting the root resorption on panoramic radiograph. We incorporate the EfficientNet for feature extraction in deep learning to the center loss and triplet loss as the loss function for metric learning. Our proposed method performed to 337 images which is obtained by panoramic radiograph. Accuracy of 71 %, true positive rate of 77%, false positive rate of 30% were obtained.

DOI： 10.23919/ICCAS52745.2021.9649745

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124211898&origin=inward

Classification of Respiratory Sounds by Generated Image and Improved CRNN 査読有り国際誌

Asatani N., Kamiya T., Mabu S., Kido S.

International Conference on Control, Automation and Systems 2021-October 1804 - 1808 2021年01月

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）

The death toll from respiratory illness reached nearly 8 million in 2019. Auscultation is used to diagnose for respiratory illness. Highly accurate diagnosis is required to reduce the number of deaths. However, unlike diagnostic imaging, auscultation of respiratory sounds could not visualize the diagnostic results. In addition, since there is a problem that the experience of a doctor affects the diagnosis results, it is required to develop a diagnostic system for quantitative analysis. In recent years, the development of a diagnostic system using the ICBHI 2017 Challenge Respiratory Sound Database has been carried out in the field of respiratory sound analysis. However, the proposed system still has accuracy problems. Therefore, in this study, we improve the proposed method by classifying the improved CRNN (Convolutional Recurrent Neural Network) by inputting multiple respiratory sound images. As a result, Sensitivity: 0.64, Specificity: 0.83, Average Score: 0.74, Harmonic Score: 0.72 were obtained, and excellent results were achieved compared with other methods.

DOI： 10.23919/ICCAS52745.2021.9649906

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124181434&origin=inward

Automatic Segmentation of Finger Bone Regions from CR Images Using Improved DeepLabv3+ 査読有り国際誌

Ono H., Murakami S., Kamiya T., Aoki T.

International Conference on Control, Automation and Systems 2021-October 1788 - 1791 2021年01月

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）

The number of hospitalized patients and the number of people requiring nursing care are serious social problems in Japan due to the increasing elderly population. The major causes of bedridden patients are bone and joint disorders caused by rheumatoid arthritis and osteoporosis. Early detection and treatment of these bone diseases are important because they significantly interfere with the quality of life (QOL) as the symptoms progress. Visual screening based on CR is used as a diagnosing tool for bone diseases. However, imaging diagnosis is subjective and lacks objectivity, and there is a possibility that lesions may be overlooked. In addition, it is difficult to find out subtle changes from images, increasing the workload for doctors. To solve these problems, there is a need to develop a computer aided diagnosis (CAD) system that can quantitatively diagnose bone diseases. We propose a method for automatic extraction of phalange regions for the CAD system to diagnose these diseases. The proposed method can extract the phalanges with high accuracy by using the improved DeepLabv3+. In this paper, we apply the proposed method to 101 cases of CR images and mIoU of 0.949 was obtained.

DOI： 10.23919/ICCAS52745.2021.9649864

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124185622&origin=inward

Automatic Identification of CTC in Fluorescence Microscope Images Using Segmentation Algorithm of Cell Nucleus 査読有り国際誌

Hashimoto K., Kamiya T., Yoneda K., Tanaka F.

International Conference on Control, Automation and Systems 2021-October 2051 - 2054 2021年01月

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）

Today, cancer is the number one cause of death in Japan, and cancer accounts for 27.3% of all deaths number. The development of cancer by repeating metastasis, hence it is important to operate early detection and early treatment. The diagnosis of cancer includes various treatments, but it is difficult to judge whether cancer is metastatic or not. Then, analysis of Circulating Tumor Cells (CTCs) in blood has been attracting attention as a new biomarker. However, because the ratio of CTCs in a billion blood cells is only a few, and there is a concern that the burden on doctors will increase. We propose a method for automatic identification of CTCs from fluorescence microscopy images to enable quantitative analysis by computer for the diagnosis of CTCs in blood. First, after detecting the cell candidate regions mainly by filtering, we set the region of interest in the cell candidate regions and reconstruct the region of interest by cutting out the cell nucleus region. In this paper, we applied the proposed method to 5,040 images of 6 samples and conducted experiments on the identification of CTCs. As a result, the number of detections was 148(TPR = 100%) and the number of over-detected non-CTCs was 988.

DOI： 10.23919/ICCAS52745.2021.9650022

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124240604&origin=inward

A Method for Evaluating of Asymmetry on Cleft Lip Using Symmetry Plane 査読有り国際誌

Sawada S., Kamiya T., Kimura-Nomoto N., Okawachi T., Nozoe E., Nakamura N.

International Conference on Control, Automation and Systems 2021-October 1796 - 1799 2021年01月

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）

Cleft lip is one of the most common morphological abnormalities that occur when the fetal lips are torn. Cleft lip is treated with multiple surgical procedures. However, there is a problem that the evaluation of surgical results depends on the subjectivity of doctors and the degree of symmetry is difficult to evaluate quantitatively. Therefore, as a previous study, a method for detecting and evaluating the plane of symmetry of the face with an asymmetric region was proposed, however, there remained a problem with the accuracy of separation of the asymmetry between patients and healthy subjects. In this paper, we propose an asymmetry evaluation method with high separation accuracy using the plane of symmetry of the face. The plane of symmetry used as the evaluation standard is obtained by adjusting the plane of symmetry of the entire face to the plane of symmetry of the surgical region. The asymmetry is evaluated by the evaluation index using the depth information with the plane of symmetry as the reference. By comparing the proposed method with the conventional method, the degree of separation of asymmetry between patients and healthy subjects was improved, and more useful results were obtained as an evaluation index.

DOI： 10.23919/ICCAS52745.2021.9649934

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124214360&origin=inward

Automatic Classification of Respiratory Sounds Based on LSTM with Affinity Loss

Minami, Kamiya, Mabu, Kido

8th International Symposium on Applied Engineering and Sciences (SAES2020) 2020年12月

A Detection Method for Liver Cancer Region Based on Faster R-CNN and Level-set Method

Furuzuki, Kamiya

8th International Symposium on Applied Engineering and Sciences (SAES2020) 2020年12月

多元計算解剖学における高度知能化診断招待有り査読有り

平野、神谷、木戸

Medical Imaging Technology 38 ( 5 ) 217 - 221 2020年11月

深層学習による胸部CT画像からの結節状陰影の検出

玉井、神谷、青木、木戸

第33回バイオメディカル・ファジィ・システム学会年次大会講演論文集 ( 第33回バイオメディカル・ファジィ・システム学会年次大会講演論文集 ) 26 - 29 2020年10月

　前のページ - 次のページ　