Machine Learning for Computer Vision
From the Boosting Idea to Deep Learning
报 告 人：Seong Kong 教授 （Sejong University）
主 持 人：赵永强 教授
报告简介：Computer vision research attempts to provide computers with human-like perception capabilities so that they can sense surrounding environments, understand the contents of sensed data, and take appropriate actions. There has been increasing demand for computer vision systems to cope with “real-world” problems. Many such applications require computer vision algorithms and systems to work under partial occlusion, highly cluttered background, low contrast, and varying illumination conditions. This requires that the vision techniques should be robust and flexible to optimize performance in a given scenario. Machine learning technology has demonstrated strong potential to contribute to the development of robust and flexible vision algorithms, thus improving the performance of practical computer vision systems. The field of machine learning is driven by the idea that computer vision algorithms can improve their future performance with time by learning from experiences. More specifically, machine learning offers effective methods for computer vision for automating model acquisition, updating processes, adapting task parameters and representations, and using the experience to generate, verify, and modify hypotheses. This talk addresses the use of several machine learning techniques into computer vision applications from the idea of boosting to deep learning. An innovative combination of computer vision algorithms and machine learning techniques is expected to be able to overcome the limitation of existing computer vision approaches. The effective use of machine learning technology in real-world computer vision problems requires the understanding of application domain, abstraction of a learning problem from a given computer vision task, and the selection of appropriate representations for the input and internal parameters of the system. Some of these aspects will be discussed from a perspective on the use of machine learning to capture the variations in visual appearances.
Seong G. Kong received the B.S. and M.S. degrees from Seoul National University, Seoul, Korea, in 1982 and 1987, respectively, and the Ph.D. degree from University of Southern California, Los Angeles, CA, USA, in 1991, all in electrical engineering. He was an Associate Professor in the Department of Electrical and Computer Engineering, University of Tennessee, Knoxville and Temple University, Philadelphia, PA, USA. He is currently a Professor of Computer Engineering at Sejong University, Seoul, Korea. His research interests include image processing, computer vision, and intelligent systems. He was a recipient of the Most Cited Paper Award from the Computer Vision and Image Understanding journal in 2007 and 2008, the Honorable Mention Paper Award from American Society of Agricultural and Biological Engineers in 2005. His professional services include as an Associate Editor of the IEEE Transactions on Neural Networks, a Guest Editor of a special issue of International Journal of Control, Automation and Systems and Journal of Sensors.