主讲嘉宾 : Prof. Yen-Wei Chen

时间： 2023年9月4日上午 10:00 – 11:00

地点： 深圳大学丽湖校区A2-517会议室

主持人： 倪东

Summary

Recently, deep learning (DL) plays important roles in many academic and industrial areas especially in computer vision and image recognition. Deep learning uses a neural network with deep structure to build a high-level feature space. It learns data-driven, highly representative, hierarchical image features, which have proven to be superior to conventional hand-crafted low-level features and mid-level features. Deep learning (DL) has also been applied to medical image analysis. Compared with DL-based natural image analysis, there are several challenges in DL-based medical image analysis due to their limited number of labeled training samples，high dimensionality and multimodality. In this talk, I will talk about several solutions for these challenges. I will first introduce deep atlas prior, in which we combined semi-supervised deep learning with anatomic atlas as prior information to solve the problem of limited annotated data. Then I will introduce VolumeNet, in which we proposed an efficient but accurate lightweight 3D network for medical volumetric data analysis. As third topic, I will introduce genotype-guided radiomics signature (GGR), in which we used gene information as a guidance for accurate CT-based recurrence prediction of lung cancer. I will also discuss futures of DL in medical imaging.

Biography

Yen-Wei Chen received the B.E. degree in 1985 from Kobe Univ., Kobe, Japan, the M.E. degree in 1987, and the D.E. degree in 1990, both from Osaka Univ., Osaka, Japan. He was a research fellow with the Institute for Laser Technology, Osaka, from 1991 to 1994. From Oct. 1994 to Mar. 2004, he was an associate Professor and a professor with the Department of Electrical and Electronic Engineering, Univ. of the Ryukyus, Okinawa, Japan. He is currently a professor with the college of Information Science and Engineering, Ritsumeikan University, Japan. He is the founder and the first director of Center of Advanced ICT for Medicine and Healthcare, Ritsumeikan University. He was a chair professor with the college of computer technology and science, Zhejiang University, China during 2014-2016.

His research interests include medical image analysis, computer vision and computational intelligence. He has published more than 300 research papers in a number of leading journals and leading conferences including CVPR, ICCV, MICCAI, IEEE Trans. Image Processing, IEEE Trans. Medical Imaging. He has received many distinguished awards including ICPR2012 Best Scientific Paper Award, 2014 JAMIT Best Paper Award. He is/was a leader of numerous national and industrial research projects.

近年来，深度学习（DL）在许多学术和工业领域发挥着重要作用，特别是在计算机视觉和图像识别领域。深度学习利用具有深层结构的神经网络构建高层特征空间。它学习数据驱动的、高度代表性的分层图像特征，这些特征已被证明优于传统的手工制作的低级特征和中级特征。深度学习（DL）也已应用于医学图像分析，与基于深度学习的自然图像分析相比，基于深度学习的医学图像分析由于其标记训练样本数量有限、高维和多模态而面临一些挑战。

在本次讲座中，我将讨论针对这些挑战的几种解决方案。我首先介绍一下深度图谱先验，其中我们将半监督深度学习与解剖图谱作为先验信息相结合，以解决标注数据有限的问题。然后我将介绍VolumeNet，其中我们提出了一种高效但准确的轻量级3D网络，用于医学体积数据分析。第三，我将介绍基因型引导的放射组学特征（GGR），其中我们使用基因信息作为基于CT的肺癌复发准确预测的指导。我还将讨论深度学习在医学成像领域的未来。

简介：

陈延伟(Yen-WeiChen)荣获学士学位1985年获得日本神户神户大学博士学位，1987年获得硕士学位， 1990年获得博士学位，均毕业于日本大阪大学。1991年至1994年任大阪激光技术研究所研究员。1994年10月至2004年3月任日本琉球大学电气电子工程系副教授、教授。现任日本立命馆大学信息科学与工程学院教授。同时是日本立命馆大学医疗保健先进ICT中心的创始人和首任主任。2014-2016年任浙江大学计算机技术与科学学院讲座教授。

他的研究兴趣包括医学图像分析、计算机视觉和计算智能。他在CVPR、ICCV、MICCAI、IEEETrans等多个领先期刊和领先会议上发表了300多篇研究论文。图像处理，IEEETrans。医学影像。他获得了许多杰出奖项，包括ICPR2012最佳科学论文奖、2014年JAMIT最佳论文奖。主持多项国家级和行业级科研项目。