研究者詳細 - 神谷　亨

論文 - 神谷　亨

分割表示751 件中 81 - 100 件目／全件表示 >>

Classification of Histological Types of Primary Lung Cancer from CT Images Using Clinical Information 査読有り国際誌

Honda N., Kamiya T., Kido S.

International Conference on Control, Automation and Systems 1753 - 1757 2023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）

Identification of primary lung cancer is very important because it influences the course of treatment, especially for small cell carcinomas, which metastasize rapidly and must be detected at an early stage. In addition to imaging, clinical information is often used in CAD (computer aided diagnosis) systems. In addition to images, clinical information is often used in CAD systems, especially information on smoking history, which is considered to be important in the diagnosis of lung cancer. In this paper, we propose a method to identify primary lung cancer by adding clinical information from medical records in addition to images in order to improve the accuracy of diagnosis. We use tumor images surrounded by rectangular regions from CT images in an open dataset as input images and train the method by deep learning. We evaluate the proposed method by discriminating tumors from unknown data. In our experiments, we found that the accuracy was improved by about 5% when clinical information was added to 655 images, which included four classes of cancer: adenocarcinoma, small cell carcinoma, squamous cell carcinoma, and large cell carcinoma.

DOI： 10.23919/ICCAS59377.2023.10316865

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179178408&origin=inward

Classification of Driver Gene Mutations from 3D-CT Images Based on Radiomics Features 査読有り国際誌

Watanabe S., Kamiya T., Terasawa T., Aoki T.

International Conference on Control, Automation and Systems 1733 - 1736 2023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）

Cancer is a leading cause of death worldwide, accounting for nearly 10 million deaths in 2020. Lung cancer is the most common cause of death from this cancer. Diagnosis is made mainly by biopsy, which checks for mutations in the driver genes of lung cancer. If there is a mutation, molecularly targeted drugs with higher therapeutic efficacy can be used. However, it is difficult for physicians to make a decision and places a heavy burden on patients. To solve this problem, a computer aided diagnosis (CAD) system is needed to identify the presence or absence of driver gene mutations from CT images. In this paper, Radiomics is used to extract features from 3D lung cancer in CT images and analyze them using machine learning to classify the presence or absence of mutations. Because the features obtained is huge, dimensionality reduction is performed using Null Importance. Furthermore, the accuracy is improved by adding gender as clinical information. The proposed method was applied to a dataset consisting of 175 cases. As a result, we obtained an AUC (Area Under the Curve) of 0.985, accuracy of 92.0%, true positive rate of 85.7%, and false positive rate of 2.20%.

DOI： 10.23919/ICCAS59377.2023.10316810

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179182718&origin=inward

Automatic Segmentation of Finger Bone Regions from CR Image Using the Improved HRNet 査読有り国際誌

Hiraoka T., Kamiya T., Aoki T.

International Conference on Control, Automation and Systems 1498 - 1501 2023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）

In Japan, it is serious problems such as increasing nursing care of handicapped people requiring and the aging of caregivers which is caused in a super-aging society. One of the main causes of being classified as persons requiring nursing care include joint diseases such as rheumatoid arthritis and bone fractures due to osteoporosis. Early detection and treatment of these diseases are considered important. Rheumatoid arthritis and osteoporosis are generally diagnosed by simple X-ray examination. However, there are problems with radiographic diagnosis by physicians, such as lack of objectivity and reproducibility of diagnosis, and increased workload on the radiologists. To solve these problems, a Computer-Aided Diagnosis (CAD) system is being developed. Because the CAD system may use the results of quantitative computer analysis, it is expected to improve the reproducibility and accuracy of diagnosis and reduce the burden on physicians. Therefore, this paper proposes a segmentation method of phalanx region from CR images to develop a CAD system. The proposed method is based on HRNet + JPU + U-Net. The proposed method was applied to 101 cases of X-ray images, and mIoU=0.897 was obtained. Experiments for segmentation from images confirmed the usefulness of the proposed method by improving the extraction accuracy of the boundary of the phalanx region.

DOI： 10.23919/ICCAS59377.2023.10317032

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179183712&origin=inward

An Image Registration Technique for Brain MR Images Using Linear Transform by 3DCNN 査読有り国際誌

Baba S., Kamiya T.

International Conference on Control, Automation and Systems 1762 - 1765 2023年01月

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 focus on brain atrophy in Alzheimer's disease and propose a 3D image registration method of brain MR images for the purpose of calculating the atrophy rate by temporal subtraction technique. The nonlinear 3D image registration model of Stergios et al. is recombined into a linear image registration model, and image registration is performed with isotropic affine and rigid body transformations. In addition, the accuracy is improved by the Instance-specific Optimization, which performs the optimization process for each case. For the training of the model, we used pseudo-data obtained by arbitrarily linear transformation of the ADNI brain MR images. Synthetic data and actual ADNI time series data (AD, MCI) were used for the test. The experimental results show that the addition of the optimization process improves the normalized cross-correlation (NCC) by more than 0.1. The loss function of the model is compared between Grid loss and MSE, and it is confirmed that MSE is effective for NCC. Furthermore, the output images confirm the usefulness of the image registration of the skull in the linear transformation.

DOI： 10.23919/ICCAS59377.2023.10316755

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179178497&origin=inward

An Denoising Method Low-Dose CT Image Using Image Restoration Convolutional Neural Network 査読有り国際誌

Sadamatsu Y., Murakami S., Guangxu L（天津工大）., Kamiya T.

International Conference on Control, Automation and Systems 1737 - 1740 2023年01月

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）

Radiation is widely used in medicine to diagnose and treat disease. CT (Computed Tomography) scans allow early detection of externally invisible diseases and appropriate treatment. However, radiation exposure from the examination may result in a future risk of cancer. Efforts are therefore being made to reduce radiation exposure. During the examination, noise is generated in the image when the dose is reduced. Noise reduces the visibility of the image and may cause lesions to be missed. In this study, we focus on Convolutional Neural Networks (CNNs), a type of deep learning model that has recorded high accuracy in image processing tasks. The proportion of frequency components in an image has more low-frequency components and fewer high-frequency components. Since image features are treated equally across channels, important information such as noise and edges are easily lost. To solve this problem, we propose CNN with channel attention module. In addition, we employ MAE as the loss function to enable effective learning. Using whole body slice CT images of pigs, we evaluate the image quality by Peak Signal-to-Noise Ratio (PSNR) and show that the proposed method is effective.

DOI： 10.23919/ICCAS59377.2023.10317050

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179178606&origin=inward

A Method for Denoising from Low-Dose CT Images Based on Noise2Noise 査読有り国際誌

Sawada S., Murakami S., Guangxu L., Kamiya T.

International Conference on Control, Automation and Systems 1494 - 1497 2023年01月

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 has the highest number of CT scanners per million people among developed countries, and the annual medical radiation dose is about six times higher than the world average. Screening of high-risk populations by means of low-dose CT has been verified to be able to reduce lung cancer mortality. However, the low-dose CT image noise should be accurately estimated in order to achieve good image quality as normal dose CT. In this paper, we apply a denoising method using convolutional neural networks without use of the clean images during learning phase. The CNN model is based on SRResNet, which has achieved high performance in super-resolution tasks, and has two attention mechanisms, Spatial Attention and Channel Attention, in the Residual Block. Experiments using whole body slice CT images of pigs showed that the proposed method is useful by comparison with normal dose CT images and by evaluation of image quality using peak signal-to-noise ratio (PSNR).

DOI： 10.23919/ICCAS59377.2023.10317004

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179178164&origin=inward

A Detection Method for Nodular Shadows from Temporal Subtraction Images Using A Machine Learning Technique Incorporated Radiomics Features 査読有り国際誌

Baba N., Kamiya T., Terasawa T., Aoki T., Kido S.

International Conference on Control, Automation and Systems 1741 - 1744 2023年01月

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）

Lung cancer progresses rapidly, and early detection and treatment are important. Computed tomography (CT) equipment is mainly used for the examination. However, the number of CT images is huge and places a heavy burden on physicians. Therefore, a computer aided diagnosis (CAD) system is necessary to overcome these problem. One of the CAD is temporal subtraction technique. The technique is that performs a subtraction operation using the current and past images of the same patient to remove normal structures and emphasize newly appearing areas that have changed over time. However, because two images taken at different dates are used, artifacts due to misalignment often occur. Therefore, it is necessary to classify partial images with high false-positive rates and enhanced temporal changes into lesions and normal tissue. In this paper, the temporal subtraction image generation method is used to extract candidate areas of abnormal shadows as initial shadows. The proposed method was applied to 22 cases of chest CT images, and the results showed that the proposed method achieved a discrimination accuracy of Accuracy=74.54, TPR=72.65, FPR=22.59, and AUC=79.16 in XGBoost, which is a kind of machine learning method.

DOI： 10.23919/ICCAS59377.2023.10316767

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179180349&origin=inward

Preface for ACPR 2023 Proceedings 査読有り国際誌

Liu C.L., Yagi Y., Kamiya T., Blumenstein M., Lu H., Yang W., Cho S.B.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 14408 LNCS v - vi 2023年01月

機械学習による胸部CT画像からのドライバー遺伝子情報変異有無の識別法

吉福，神谷，寺澤，青木

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 2022年12月

Recognition of sidewalk environment based on WideSegPlus 査読有り国際誌

Sakai, Lu, Yujie Li, Kamiya

Proceedings Volume 12508, International Symposium on Artificial Intelligence and Robotics 2022 ( SPIE ) 2022年12月

Grasping position estimation method using depth image for thin objects 査読有り国際誌

Yoshihara, Koga, Lu, Kamiya

Proceedings Volume 12508, International Symposium on Artificial Intelligence and Robotics 2022 ( SPIE ) 2022年12月

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

Horimachi, Lu, Zheng, Kamiya, Nakatoh, Serikawa

Proceedings Volume 12508, International Symposium on Artificial Intelligence and Robotics 2022 ( SPIE ) 2022年12月

Underwater video networking and targe tracking 査読有り国際誌

Ota, Lu, Jianru Li, Yuchao Zheng, Kamiya, Nakatoh, Serikawa

Proceedings Volume 12508, International Symposium on Artificial Intelligence and Robotics 2022 2022年12月

Automatic Classification of Respiratory Sound Considering Hierarchical Structure 査読有り国際誌

Tabata M., Lu H., Kamiya T., Mabu S., Kido S.

International Conference on Control, Automation and Systems 2022-November 537 - 541 2022年11月

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）

Respiratory diseases are one of the leading causes of death worldwide. Approximately 8 million people die annually from respiratory diseases. Diagnosis is made primarily by auscultation using a stethoscope. The lack of quantitative criteria makes diagnosis difficult in the field where physicians are in short supply. To solve this problem, a computer aided diagnosis (CAD) system that quantitatively analyzes and classifies respiratory sounds and outputs them as a 'second opinion' is needed. In this paper, HPSS (Harmonious/Percussive Sound Separation) is used to separate abnormal respiratory sound features. Images are generated from the spectral envelopes obtained by linear prediction coefficients (LPC) for each of the three types of respiratory sound data before separation. The CNN (convolutional neural networks) framework based on hierarchical structure of the correct labels is introduced. The proposed method was applied to the dataset used in the International Conference on Biomedical and Health Informatics (ICBHI) 2017 Challenge. As a result, we obtained a sensitivity of 63.5%, specificity of 85.1%, average score of 74.3%, harmonic score of 72.7%, area under the curve of 87.8%, and false negative rate of24.5%, respectively.

DOI： 10.23919/ICCAS55662.2022.10003771

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85146537757&origin=inward

Automatic Identification of CTCs in Fluorescence Microscope Images Using Morphological Filtering to Detect Cell Nuclei 査読有り国際誌

Hashimoto, Park, Ha, Jung, Kamiya

International Conference on Control, Automation and Systems 554 - 557 2022年11月

Automatic Detection of Cervical Lymph Nodes from Non-Contrast CT Images Using the SSD 査読有り国際誌

Baba, Ishida, Kamiya

International Conference on Control, Automation and Systems 566 - 569 2022年11月

Recognition of Plastic Bottle Using Improved U-Net 査読有り国際誌

Ideta D., Kamiya T.

International Conference on Control, Automation and Systems 2022-November 200 - 203 2022年11月

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 focuses on separation rubbish, which is one of the causes of global warming. Currently, recyclable waste including garbage is sent to the waste disposal site in an unsorted state, and it is separated by human source. This is difficult work and a cause of manpower shortage. To overcome those problem, it is necessary to introduce automatic sorting algorithms on separation rubbish process. In this study, we focused on separation rubbish process especially plastic bottles based on image recognition technique using a deep learning approach. We implement an improved U-Net as the deep learning scheme to increase segmentation accuracy. As a result, we obtained the segmentation accuracy with 0.789 of cap part and 0.972 of body part on plastic bottle respectively.

DOI： 10.23919/ICCAS55662.2022.10003819

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85146598114&origin=inward

Identification of abnormal tissue from CT images using improved ResNet34 査読有り国際誌

Honda N., Kamiya T., Kido S.

International Conference on Control, Automation and Systems 2022-November 532 - 536 2022年11月

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, CT examinations have been widely used as a screening method to detect lung cancer. However, reading enormous CT images become a heavy burden to the physician. To avoid this problem, computer-aided diagnosis systems have been introduced on CT screening. In general, physicians consider patient information in addition to image information when they make a diagnosis, new efforts are being made to improve the accuracy of diagnosis by mimicking this information with a machine. In this paper, we propose a method for identifying pulmonary nodules by adding medical record information to images to improve the accuracy of diagnosis. We classify nodules from unknown data by assigning branching information of vascular opacities, straight vascular shadows, and nodular shadows as labeled image, which are a cause of misrecognition based on image features in machine learning. In the experiment, the classification accuracy of the nodule class was improved by adding clinical information to 644 images including 161 nodal images.

DOI： 10.23919/ICCAS55662.2022.10003937

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85146569950&origin=inward

Identification for Drinking Spout of Plastic Bottle Using Multi Eff-UNet 査読有り国際誌

Moritsuka S., Kamiya T.

International Conference on Control, Automation and Systems 2022-November 209 - 212 2022年11月

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）

Currently, the increase of general waste is a problem in Japan. It is garbage and human waste. One solution to this problem is to reuse the bodies of used plastic bottles to reduce the amount of plastic waste and the increase in general waste. For this purpose, some municipalities provide bags for disposing of plastic bottle bodies. However, in many cases, the bottles are not only discarded in these bags, but also with their caps attached. In such cases, the caps must be found and removed by hand in the sorting and processing facility at the waste treatment plant. This means that the same work is performed by person for a long time, causing problems such as overlooking caps due to fatigue. To solve these problems, we develop a method for automatic identification of plastic bottle bodies and caps using deep learning technique. In this paper, we propose a model that combines multiple Eff-UNets. Specifically, we combine EfficientNetB4 for local segmentation and EfficientNetB5 for global segmentation. By using our method, we conducted an experiment on images of plastic bottles collected from the internet and other sources. We obtained segmentation results of 97.9% for plastic bottle bodies and 86.0% for plastic bottle caps.

DOI： 10.23919/ICCAS55662.2022.10003920

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85146540644&origin=inward

Generation of Super-Resolution Images from Satellite Images Based on Improved RCAN 査読有り国際誌

Morishima F., Lu H., Kamiya T.

International Conference on Control, Automation and Systems 2022-November 213 - 216 2022年11月

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）

Satellite images can be analyzed and used for a variety of purposes. In the future, satellite image analysis will become more important since the number of satellites launches, and the amount of satellite data increase every year. Under these circumstances, there are some problems to be solved. One is the existence of low-resolution satellite images. To analyze the lower resolution of satellite images there are some technical issues such as reduction of noise, misclassification of object recognition. Therefore, high-resolution images are necessary. However, high-resolution satellite images are expensive, and its images may not be available in the past satellite images. Super-resolution which is introduced in image processing is a method to solve these problems. Convolutional neural network (CNN)-based methods are effective, and there is a need for models that can perform super-resolution with higher accuracy. In this paper, we propose a method for super-resolving satellite images, based on the improved the RCAN (residual channel attention network) model with SRM (style-based recalibration module). The proposed method improves the PSNR (peak signal to noise ratio) by 0.0181 dB compared to the conventional RCAN model.

DOI： 10.23919/ICCAS55662.2022.10003856

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85146557504&origin=inward

　前のページ - 次のページ　

<神谷 亨>

論文 - 神谷 亨

<神谷　亨>

論文 - 神谷　亨