Recognizing Human Actions Seen in Video
Dr. Dipti Prasad Mukherjee
5 de Octubre de 2015
12:00 hrs.
Auditorio de Ingeniería Eléctrica.
|
Texto completo de la plática 
Resumen
We propose a graph theoretic technique for recognizing human actions (for example, running, walking, jumping etc.) seen in a video. The recognition of actions is based on a model of the visual senses associated with human poses. Identifying the intended meaning of poses is a challenging task because of their variability. Such variations in poses lead to visual sense ambiguity. Our methodology follows a bag-of-words approach. Here "word" refers to the pose descriptor of the human figure corresponding to a single video frame and a "document" corresponds to the entire video of a particular action. From a large vocabulary of poses we prune out ambiguous poses and extract key poses for each action type in a supervised fashion using the centrality measure of graph connectivity. The number of key poses per action is determined by setting a "meaningful" bound on the centrality measure. We evaluate our methodology on four standard activity recognition datasets. The results clearly demonstrate the superiority of our approach when compared to the present states-of-the art.
Breve semblanza biográfica
Dipti Prasad Mukherjee is a Professor of the Electronics and Communication Sciences Unit, Indian Statistical Institute, Kolkata. His primary research interest is in Computer Vision and Image Processing. He has written two books on Computer Graphics, edited four books and written more than one hundred peer-reviewed research papers. He had held visiting faculty positions at the Oklahoma State University and University of Virginia, USA and the University of Alberta, Canada. Prior to this, Dr. Mukherjee is the recipient of the pre-doctoral UNDP fellowship at the Robotics Research Group, University of Oxford, U.K., and the UNESCO-CIMPA fellowship to INRIA, France and to ICTP Italy. In 2010, he has received Japan Society of Promotion of Science Invitation fellowship to the Department of Radiology, Osaka University, Japan. He is a senior member of IEEE and Fellow of the Computer Society of India and the Institution of Engineers (India). He is currently an Associate Editor of IEEE Transactions on Image Processing and the Journal of the Indian Academy of Sciences, Sadhana published by Springer. He had served on the Editorial Board of the IEEE Signal Processing Letters and ISRN Machine Vision journal.
