Visit ShanghaiTech University | 中文 | How to find us
HOME > News and Events > Events
Medical Image Analysis with Deep Learning: Applications of Semantic Scoring, Anatomical Decomposition, and Image Quality Assessment
Date: 2018/5/25             Browse: 76

Speaker:     Dr. Jie-Zhi Cheng

Time:          15:10—16:00, May  25

Location:    Room 1A-200, SIST Building

Abstract:

Great progress of medical image analysis has recently been made with the deep learning techniques for the advantages of automatic feature extraction and end-to-end training. With proper network training as well as useful transferred model initialization, promising image analysis results can be achieved. In this talk, I will discuss about our recent deep learning based medical image analysis works. The specific applications include the automatic semantic scorings for CT pulmonary nodules, breast anatomical layer segmentation in whole breast ultrasound, and the image quality evaluation for fetal ultrasound. In these works, we explore the techniques of deep convolutional neural network, stacked denoising autoencoder, multi-task regression, convolutional encoder-decoder network, domain transfer, etc., to attain satisfactory performance on each application. 

Bio:

Jie-Zhi Cheng received his B.S., M.S., and Ph.D. degrees from National Taiwan University, Taipei, Taiwan, in 2002, 2007, and 2013, respectively. His undergraduate major is computer science whereas the two post-graduate degrees are both biomedical engineering. He was an Associate Professor in Department of Biomedical Engineering, School of Medicine, Shenzhen University and now is a R&D VP of United Imaging Intelligence. His research interests mainly include medical image analysis, computer-aided diagnosis and intervention, pattern recognition, and machine learning. Dr. Cheng has coauthored more than 40 scientific articles on several medical image analysis journals of IEEE TMI, Radiology, Medical Physics, Ultrasound in Medicine and Biology, etc and first tier medical image conferences like IPMI and MICCAI.

SIST-Seminar 18040