Title: Mining Middle-Level Representations for Complex Human Action Recognition

Reporter: Dr. Yu Qiao (Professor, SIAT, CAS)

Time:  15:00-16:00, 15 Jan. 2014

Address:Room 3-202, Building No. 5, School of Remote Sensing and Information Engineering, Wuhan University



Human action recognition is receiving extensive research interests in computervision nowadays due to its wide applications in surveillance, human-computerinterface, sports video analysis, and content based video retrieval. Thechallenges of action recognition come from background clutter, viewpoint changes,and motion and appearance variations.  Inthis talk, we will show our two recent works (CVPR13, ICCV13) to address thesechallenges. Both of them develop middle level parts for action representation,but different in how to define and mine the parts. Experimental results onlarge public datasets demonstrate the effectiveness of our approach. Our methodachieves the best known performance (87%) on ICCV THUMOS'13 Action RecognitionChallenge. 


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 theUniversity of Tokyo from 2007 to 2010. Now he is a professor and a CAS BaiRenscholar with the Shenzhen Institutes of Advanced Technology, the ChineseAcademy of Sciences. His research interests include pattern recognition, computervision, multimedia, image processing and machine learning. He has publishedmore than 100 papers in these fields. He received the Lu Jiaxi young researcheraward from the Chinese Academy of Sciences.