http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Detection of Lung Nodules from Temporal Subtraction Image Using Deep Learning
Kohei TAMAI,Noriaki MIYAKE,Humin LU,Hyoungseop KIM,Seiichi MURAKAMI,Takatoshi AOKI,Shoji KIDO 제어로봇시스템학회 2019 제어로봇시스템학회 국제학술대회 논문집 Vol.2019 No.10
In recent years, the number of death due to lung cancer is increasing year by year worldwide. Early detection and early treatment of lung cancer are important. Especially, early detection of the abnormalities on thoracic MDCT images detection of small nodules is required in visual screening. Although a CT apparatus is used for the examination, the burden on the image interpretation doctor is large due to the high performance of the CT, so the diagnostic accuracy may be reduced. In this paper, we propose an image analysis method to detect abnormal shadows from chest CT images automatically. The initial lesion candidate areas are extracted by using temporal subtraction technique that emphasizes temporal change by subtracting from a current image to previous one which is obtained same subject. The image of the area is given as input and classification is performed by CNN (Convolutional Neural Network). In the discrimination experiment based on our proposed method, 90.26 [%] of true positive rates and 13.58 [%] of false positive rates are obtained from the 49 clinical cases.
Yuichiro Koizumi,Noriaki MIYAKE,Huimin Lu,Hyoungseop Kim,Seiichi MURAKAMI,Takatoshi AOKI,Shoji KIDO 제어로봇시스템학회 2018 제어로봇시스템학회 국제학술대회 논문집 Vol.2018 No.10
In recent years, the proportion of deaths from cancer tends to increase in Japan, especially the number of deaths from lung cancer is increasing. CT device is effective for early detection of lung cancer. However, there is concern that an increase in burden on doctors will be caused by high performance of CT improving. Therefore, by presenting the “second opinion” by the CAD system, it reduces the burden on the doctor. In this paper, we develop a CAD system for automatic detection of lesion candidate regions such as lung nodules or ground glass opacity (GGO) from 3D CT images. Our proposed method consists of three steps. In the first step, lesion candidate regions are extracted using temporal subtraction technique. In the second step, the image is reconstructed by sparse coding for the extracted region. In the final step, 3D Convolutional Neural Network (3D-CNN) identification using reconstructed images is performed. We applied our method to 51 cases and True Positive rate (TP) of 79.81 % and False Positive rate (FP) of 37.65 % are obtained.
Detection of Abnormal Candidate Regions on Temporal Subtraction Images Based on DCNN
Mitsuaki NAGAO,Noriaki MIYAKE,Yuriko YOSHINO,Huimin LU,Joo Kooi TAN,Hyoungseop KIM,Seiichi MURAKAMI,Takatoshi AOKI,Yasushi HIRANO,Shoji KIDO 제어로봇시스템학회 2017 제어로봇시스템학회 국제학술대회 논문집 Vol.2017 No.10
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.
Detection of Abnormal Shadows on Temporal Subtraction Images Based on Multi-phase CNN
Mitsuaki NAGAO,Noriaki MIYAKE,Yuriko YOSHINO,Huimin LU,Hyoungseop KIM,Seiichi MURAKAMI,Takatoshi AOKI,Shoji KIDO 제어로봇시스템학회 2018 제어로봇시스템학회 국제학술대회 논문집 Vol.2018 No.10
Recently, visual screening based on CT images become useful tools in the medical fields. However, due to the large number of images and the complexity of the image processing algorithms, image processing technique for the high screening quality is still required. To overcome this problem, some computer aided diagnosis (CAD) algorithms are proposed. 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 and early treatment. We have designed and developed a framework combining machine learning based on multi-phase convolutional neural networks (CNN) and temporal subtraction techniques based on non-rigid image registration algorithm. Our main classification method can be built into three main steps; i) preprocessing 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 93.55%, false positive rates of 10.93 /case.
Extraction of GGO Regions from Chest CT Images Using Deep Learning
Kazuki HIRAYAMA,Noriaki MIYAKE,Huimin LU,Joo Kooi TAN,Hyoungseop KIM,Rie TACHIBANA,Yasushi HIRANO,Shoji KIDO 제어로봇시스템학회 2017 제어로봇시스템학회 국제학술대회 논문집 Vol.2017 No.10
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].