Prof. Nizhuan WANG, Jiangsu Ocean University, China
Title: Deep Networks & Its Applications: From Convolutional Sparse Coding to Graph Convolutional Network
Dr. Nizhuan Wang currently is research associate professor of School of Biomedical Engineering, ShanghaiTech University, Shanghai, China. Before that he was a full professor of School of Computer Engineering, the Director of Artificial Intelligence & Neuro-informatics Engineering (ARINE) Laboratory, and the Director of Department of Software Engineering (2019.08-2021.07), Jiangsu Ocean University from Mar. 2018 to Dec. 2021. Also, he was an assistant professor and the deputy director (Medical Information Engineering Department) of Shenzhen University (2016.06-2018.03). He received the degrees of B.S., M.S. and Ph.D. from Heilongjiang University (in 2010), Shanghai Maritime University (in 2012 and 2016), respectively. He is a senior member of Chinese Biomedical Engineering Society, a member of Chinese Artificial Intelligence Society, a member of OHBM, the Program Committee member of 2018 IEEE ICIA, the Session Chairs of 2017 IEEE ICIA and 2018 ICIS, the Keynote Speaker of 2020 ICMLCA, the Chairman and Keynote Speaker of 2021 ICBIC, Guest Associate Editor of Frontiers in Neuroscience. His master dissertation was awarded “Shanghai Outstanding Master Thesis” in 2014. At his Ph.D. study stage, he was awarded “China National Scholarship for Distinguished Doctoral Students” twice in 2013 and 2015 respectively.
Prof. Linlin Shen, School of Computer Science, Shenzhen University, China
Title: Deep Learning based Medical Image Analysis: Data Problems and Solutions
While deep learning has been widely applied in medical image analysis like lung nodule detection, brain tumor segmentation and pathology image analysis, data insufficiency and labelling costs are always two major challenges. In this talk, an overview about our research progress on deep learning based medial image analysis will be introduced. In addition to general data augmentation, tangled neuron synthesis using single neuron data and GAN based tongue image synthesis will also be presented. Due to the high cost of labelling, the number of well-labeled medical training data available for training is very limited. While a pretrained classifier can be used to provide pseudo label for the following training, self-supervised learning has also been a popular approach to mitigate the requirement of labeled data. Possible extensions of such methodologies and future directions will also be discussed.
Prof.Anhui Liang, Shandong University of Science and Technology, China
Title: Optical Fiber Characteristics of Chinese Meridian Systems and Mussel Cell Lasers
Prof. Yuanchang Zhong, Chongqing University, China
Title: Research and efficacy of embedded intelligent neural pacemaker
It has become possible to establish an artificial channel for neural action potentials, to realize the purpose of neural function recovery and reconstruction, and to get more and more attention and research.Based on the technical difficulties of high-fidelity reproduction of ultra-low-frequency neural signal, a solution based on advanced embedded intelligent chip and deep learning technology is proposed.Under the control of the intelligent chip, relying on the analog-to-digital conversion and digital-mode conversion circuit, the acquisition, digital and high-fidelity reproduction are realized.