Time:13-11-2013 9:30am
Place:Room 402, State Key Laboratory of CAD & CG, Library and Information Center Building B, Zhejiang University Zijin’gang Campus
Title of the report 1:Deep Learning in Object Detection, Segmentation and Recognition
Reporter:Dr. Xiaogang Wang
Title of the report 2:Mining Middle Level Representations for Complex Human Action Recognition
Reporter:Dr. Yu Qiao
Title: Deep Learning in Object Detection, Segmentation and Recognition
Abstract: Deep learning has become a major breakthrough in artificial intelligence and achieved amazing success on solving grand challenges in many fields including computer vision. In this seminar, I will introduce our recent works on developing deep models to solve several computer vision problems, including pedestrian detection, facial keypoint detection, face parsing, pedestrian parsing, face recognition, and face attribute recognition. Deep models significantly advance the state-of-the-art on these challenges because of their capability of automatically learning hierarchical feature representations from data, jointly optimizing key components in a computer vision system, and their learning capacity. Through examples, I will share our experience on how to formulate a vision problem with deep learning, how to train a deep neural network, how to make use the large learning capacity of deep models, and how to learn features in a vision application. The benefits of deep architectures will also be discussed.
Bio: Xiaogang Wang received his Bachelor degree in Electrical Engineering and Information Science from the Special Class of Gifted Young at the University of Science and Technology of China in 2001, M. Phil. degree in Information Engineering from the Chinese University of Hong Kong in 2004, and PhD degree in Computer Science from Massachusetts Institute of Technology in 2009. He is an assistant professor in the Department of Electronic Engineering at the Chinese University of Hong Kong since August 2009. He received the Outstanding Young Researcher in Automatic Human Behaviour Analysis Award in 2011, Hong Kong RGC Early Career Award in 2012, and Young Researcher Award of the Chinese University of Hong Kong. He is the associate editor of the Image and Visual Computing Journal. He was the area chair of ICCV 2011, ECCV 2014 and ACCV 2014. His research interests include computer vision, deep learning, crowd video surveillance, object detection, and face recognition.
Title: Mining Middle Level Representations for Complex Human Action Recognition
Abstract: Human action recognition is receiving extensive research interests in computer vision nowadays due to its wide applications in surveillance, human-computer interface, sports video analysis, and content based video retrieval. The challenges of action recognition come from background clutter, viewpoint changes, and motion and appearance variations. In this talk, we will show our two recent works (CVPR13, ICCV13) to address these challenges. Both of them develop middle level parts for action representation, but different in how to define and mine the parts. Experimental results on large public datasets demonstrate the effectiveness of our approach.
Bio: Yu Qiao received Ph.D from the University of Electro-Communications, Japan, in 2006. He was a JSPS fellow and then a project assistant professor with the University of Tokyo from 2007 to 2010. Now he is a professor and a CAS BaiRen scholar with the Shenzhen Institutes of Advanced Technology, the Chinese Academy of Sciences. His research interests include pattern recognition, computer vision, multimedia, image processing and machine learning. He has published more than 90 papers in these fields. He received the Lu Jiaxi young researcher award from the Chinese Academy of Sciences.