时间：2019年10月23日
举办单位：大连理工大学-立命馆大学智能计算医疗健康联合研究中心
会议主题：本次研讨会由大连理工大学-立命馆大学智能计算医疗健康联合研究中心举办，着重探讨人工智能、深度学习等技术在医学图像处理和分析方面的研究内容和具体应用。本次研讨会邀请日本大阪大学的木戸尚治教授和日本山口大学的间普真吾副教授给我校开发区校区的师生进行演讲，介绍其基于人工智能的肺部影像诊断方面的研究工作和科研成果，并与校内同行一起对共同感兴趣的问题深入交换看法，探寻未来可能的合作研究，同时也给相关年轻学者和博士生一个了解智能医疗计算方面的最新研究成果的机会。
日程安排：
10月23日 大连理工大学开发区校区
10:00 - 11:30 国际信息与软件学院科研情况介绍（信息楼304）
14:00 – 14:45 木戸尚治 （教学楼A-216）
14:45 - 15:30 间普真吾（教学楼A-216）
15:45 – 16:30 与国际信息与软件学院教师及同学座谈（教学楼A-216）
演讲信息
演讲人： 木戸尚治（Shoji Kido），大阪大学医学系研究科 教授
演讲题目： Current status and issues for application of artificial intelligence on diagnostic imaging for pulmonary images
演讲摘要： For assisting radiologists’ diagnoses of various kinds of lung abnormalities such as lung nodules and diffuse lung diseases, computer-aided diagnosis (CAD) systems include three types of algorithms such as “detection” that find abnormal lesions, “classification” that differentiate abnormal lesions into benign or malignant, or histological subtypes, and “segmentation” that extract organs or abnormal areas. In usual CAD algorithms, designing an image-feature extractor is important. However, this task is difficult for medical engineers. On the other hand, a CAD algorithm by use of convolutional neural network (CNN) does not require the image-feature extractor. In this presentation, I introduce an image-based CAD algorithm for classification (differential diagnosis) of lung abnormalities by use of CNN. Moreover, I introduce a computer-aided detection algorithm by use of R-CNN, and computer-aided segmentation algorithms by use of fully convolutional network (FCN) and U-net.
演讲人介绍： Shoji Kido received his M.D. degree from Osaka University in 1988. He received his Ph.D. degrees in Medicine and Information Science from Osaka University in 1992 and 1999, respectively. He was a professor in the Department of Computer Science and Systems Engineering at Yamaguchi University from 1999 to 2006, in the Department of Applied Medical Engineering Science, Graduate School of Medicine, Yamaguchi University from 2006 to 2016, and in the Division of Information Sciences and Engineering, Graduate School of Sciences and Technology for Innovation, Yamaguchi University from 2016 to 2019. He is currently a professor in the Department of Artificial Intelligence Diagnostic Radiology, Osaka University Graduate School of Medicine.
演讲人： 間普真吾（Shingo Mabu），山口大学大学院创成科学研究科，副教授
演讲题目：Computer Aided Diagnosis for Chest CT Images Using Deep Learning
演讲摘要：Deep Learning has been actively applied to computer-aided diagnosis (CAD) of medical images. One of the reasons for the success of deep learning is the availability of big data, but basically, deep learning is based on supervised learning which requires a large number of data annotated by domain experts. In the case of medical images, it is quite tough work for radiologists to annotate thousands of images for deep learning. In addition, if a CAD system trained at a certain hospital is used at other hospitals, the same performance may not be obtained due to the environmental differences, e.g., different devices, different settings, etc. To solve the above problems, we are studying algorithms that can work well even when the number of annotated data is small and environmental conditions change. In this presentation, unsupervised learning, data augmentation using Generative Adversarial Network (GAN), and data standardization using Cycle GAN are introduced to show their effectiveness in medical image analysis.
演讲人介绍：Shingo Mabu received Ph.D. degree from Waseda University, Japan in 2006 From 2006 to 2012, Assistant Professor at Waseda University. From 2012 to 2017, Assistant Professor at Yamaguchi University, Japan. From 2017, Associate Professor at Yamaguchi University. His research interests include AI, machine learning and data mining and their applications to medical data, satellite images, microscope images and so on.
