题 目：Video Compression and Analysis: Challenges and Chance in Big Data Era

主讲人：Dr. Xinfeng Zhang

时 间：2017年3月20日（周一）下午13:30-15:00

地 点：北京大学计算机所大楼　106报告厅

Abstract:

With the coming of big data era, there are more and more chances and challenges in visual signal compression and processing. Although the latest video coding standard HEVC achieves around 50% bitrate saving compared with H.264/AVC at the same quality level, it still cannot satisfy the demand for the explosive increase of big video data. In this talk, some new solutions from both encoding and decoding perspectives are introduced to alleviate the storage and bandwidth burdens for big video data. At the decoder sider, the multiple image/video prior models with their reliability are formulated to restore high quality image/video from compressed streams, which achieve the bandwidth saving with standard compatible. At the encoder side, the nonlocal reference techniques are investigated, which promote the video coding performance by reducing the nonlocal redundancy, even the set redundancy among cloud similar data. Especially, the well-designed Nonlocal Adaptive Loop Filter (NALF) improves the compression performance significantly, more than 5% bitrate saving, by utilizing the image nonlocal similar patches to construct the low-rank matrix constraint. Some more challenge problems for cloud based compression will also be discussed in this talk.

Biography:

Dr. Xinfeng Zhang received the Ph.D. degree in computer science from the Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China, in 2014. From 2009 to 2014, he worked in Institute of Digital Media, Peking University, as a visiting student. He is currently a Research Fellow in Nanyang Technological University, Singapore. His research interests include image and video compression, processing and quality assessment. He is also an active participant in video coding standard including H.265/HEVC, IEEE-1857 and AVS. He has published more than 40 journal and conference papers including IEEE Transactions on Image Processing, IEEE Transactions on Circuits and Systems for Video Technology，IEEE Transactions on Multimedia, and submitted 16 patents with 7 authorized patents.