Zhang, Chao
Associate Professor
Research Interests: Digital image processing, machine learning, computer vision
Office Phone: 86-10-6275 7000
Email: c.zhang@pku.edu.cn
Zhang, Chao received the BEng in Automatic Control, MEng in Signal Processing and Ph.D. degrees in Signal and Information Processing from Beijing Jiaotong University, Beijing, China, in 1984, 1989 and 1995, respectively. From 1995 to 1997, he was a post-doctoral research fellow with the National Laboratory on Machine Perception, Peking University. Since 1997, he has been an associate professor with the Key Laboratory of Machine Perception, School of Electronics Engineering and Computer Science, Peking University. His research interests include image processing, machine learning, and visual recognition.
Dr. Zhang has published more than 60 research papers, including the papers published in top-tier conferences and journals, such as ICCV, CVPR, AAAI, and IEEE TIP, IEEE TNNLS. He won 6 national invention patents, some of which have been applied in the video surveillance of national high-speed railway and traffic management. He has served in the Program Co-Chair of international conferences of MPR2008 and WCVIM2009, and Program Committee Member of ICHB 2011. Zhang was a recipient of a 2nd Class Award for Advancement of Science and Technology ( Ministry of Education,1997), Outstanding Award for Achievement of Science and Technology (Ministry of Labor and Social Security, 2003), Chia Tai Teaching Award (Peking University, 2004) and Huawei Teaching Award (Peking University, 2007).
Dr. Zhang has more than ten research projects funded by NSFC, 973 programs, Ministry of Railway, and Ministry of Labor and Social Security, etc. His research achievements are summarized as follows:
1) Algorithms and theories for low-rank model in machine learning: Due to an explosion of big data from all fields of science, there is an increasing need for computational and statistical techniques and tools for data analytics, recognition, learning with relatively limited samples in high dimensions. Based on the low-rank model, he proposed some new algorithms for robust subspace recovery, dictionary learning, image and video analyses with simultaneous feature selection and extraction.
2) The key algorithms of intelligent surveillance systems: Investigate online semi-supervised learning algorithms for the key components of video surveillance systems, including background subtraction, motion detection, object tracking and behavior analysis. By marked a small number of false alarmed samples, online video analysis algorithms will update as combined decision of the given priors and on-line classifiers, thus alleviating the human intervention and improve the accuracy of the algorithm. The key algorithms have been used in the intelligent video surveillance systems of high-speed railway transportation.
3) Large-scale fingerprint identification and retrieval: Based on theories and techniques of machine learning, he explored the feature representations of fingerprint and developed new identification and fast retrieval methods. The research of project greatly improved the performance of the original automatic fingerprint identification system. The new techniques have been used in the fingerprint authentication systems for national pension payment.
