【15-28期VALSE Webinar活动】 报告嘉宾：李永杰（电子科技大学）
主持人：彭玺（A*STAR）
报告时间：2015年9月2日周三晚20:00（北京时间）
报告题目：Image Processing Models Inspired by Two Kinds of Double-Opponent Neurons in the Primary Visual Cortex [Slides]
文章信息：
[1] Kaifu Yang, Shaobing Gao, Chaoyi Li, Yongjie Li*, Efficient Color Boundary Detection with Color-opponent Mechanisms, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2013, pp2810-2817.
[2] Shaobing Gao, Kaifu Yang, Chaoyi Li, Yongjie Li*, A Color Constancy Model with Double-Opponency Mechanisms, IEEE International Conference on Computer Vision (ICCV), 2013, pp929-936 (oral paper).
[3] Shaobing Gao, Kaifu Yang, Chaoyi Li, Yongjie Li*, Color Constancy Using Double-Opponency, IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015 (accepted). (DOI: 10.1109/TPAMI.2015.2396053)
[4] Kai-Fu Yang, Shao-Bing Gao, Ce-Feng Guo, Chao-Yi Li, and Yong-Jie Li*, Boundary Detection Using Double-Opponency and Spatial Sparseness Constraint, IEEE Transactions on Image Processing, 2015, 24(8): 2565-2578.
报告摘要：In this talk I will try to explain the functional roles of several types of color-opponent cells in the primary visual cortex (V1) from the computational view of point. In particular, I will talk about a computational framework based on the color-opponent mechanisms of a certain type of color sensitive double-opponent (DO) cells in V1. This type of DO cells has oriented receptive field (RF) with both chromatically and spatially opponent structure. Experimental results show that this type of DO cells can flexibly capture both the structured chromatic and achromatic boundaries of salient objects in complex scenes when the cone inputs to DO cells are unbalanced. In addition, I will also talk about another computational framework by simulating another type of DO cells in V1 whose RFs have concentrically organized center-surround structure and are both spectrally and spatially opponent. Experimental results show that the color distribution of the responses of such type of DO cells to the color-biased images coincides well with the vector denoting the light source color. Then the illuminant color can be easily estimated by pooling the responses of DO cells. With the illuminant estimate, color constancy is easily achieved for the color-biased images.
报告人简介：Yongjie Li (李永杰) is currently a professor with the School of Life Science and Technology, University of Electronic Science and Technology of China (UESTC, 电子科技大学). He was a visiting scholar at Dept of Neuroscience, Columbia University during Sept, 2009 ~ Sept, 2010. He is a principal investigator of 4 research projects sponsored by NSFC of China and 973 Projects. Before joining UESTC, He received his Ph.D. Degree in biomedical engineering from UESTC, Chengdu, China, in June 2004. His current research interest focuses on the modeling of the information processing mechanisms of biological vision and its applications in computer vision (e.g., image/video processing). He has so far published more than 60 international journal papers (e.g., Neuroimage, Frontier in Neural circuits, IEEE Trans PAMI, IEEE Trans IP, IEEE Trans BME), conference papers (e.g., ICCV/CVPR/ECCV, including 1 ICCV oral paper) and book chapters, covering such topics as modeling of light/dark adaptation, non-classical receptive fields, color constancy, selective attention, etc., and their applications in image enhancement, denosing, dehazing, high-dynamic range image rendering, color correction, contour detection, image segmentation, saliency detection, etc. He holds 17 Chinese patents (10 granted, 7 pending). |
