Talk on Dictionaries, Manifolds and Domain Adaptation Methods for Image and Video-based Recognition

Dear All,

The Department of Electrical Engineering, the IEEE Signal Processing Society, Bangalore Chapter, and the IEEE Bangalore Section cordially invite you to a talk on

Title : Dictionaries, Manifolds and Domain Adaptation Methods for Image and Video-based Recognition

Speaker : Prof. Rama Chellappa
Minta Martin Professor of Engineering
Director, Center for Automation Research
University of Maryland, College Park, MD 20742, USA

Date :  Wednesday 19th December 2012, 3:30pm
Venue : MMCR, Department of Electrical Engineering

Talk Abstract :

Feature extraction or representation of patterns and adaptation of classifiers designed using training data to be effective on testing data are two fundamental problems in pattern recognition. In this talk, I will discuss new solutions to these problems based on theories of dictionary learning, analytic manifolds and domain adaptation with applications in image and video-based recognition. Specifically, I will discuss methods for representing images and videos using linear and non-linear dictionaries and analytical manifolds. I will then discuss methods for adapting the dictionaries and manifold representations for addressing shifts in data distributions due to changes in pose, illumination, spatio-temporal sampling and blur with applications in recognition of faces, expressions, objects and actions.

Speaker Biography :

Prof. Rama Chellappa received an ME degree in Electrical Communication Engineering from the Indian Institute of Science in 1977. Subsequently he received the M.S.E.E. and Ph.D. Degrees in Electrical Engineering from Purdue University in 1978 and 1981 respectively. During 1981-1991, he was an Assistant (1981-1986) and Associate Professor (1986-1991) and Director of the Signal and Image Processing Institute (1988-1990) at University of Southern California, Los Angeles. Since 1991, he has been a Professor of Electrical and Computer Engineering (ECE) and an affiliate Professor of Computer Science at University of Maryland (UMD), College Park.  He is also affiliated with the Center for Automation Research  Director), the Institute for Advanced Computer Studies (Permanent Member) and is serving as the Chair of the ECE department.  In 2005, he was named a Minta Martin Professor of Engineering. His current research interests are face recognition, clustering and video summarization, 3D modeling from video, image and video-based recognition of objects, events and activities, dictionary-based inference, compressive sensing, domain adaptation and hyper spectral processing.

Prof. Chellappa has received several awards, for teaching, research, service from USC, UMD, IAPR, IEEE Signal Processing and Computer Societies. He was awarded the 2012 King-Sun Fu Prize by the International Association for Pattern Recognition for pioneering contributions to statistical methods for image and video-based object recognition. In 2010, he was recognized as an Outstanding ECE by Purdue University. He is a Fellow of IEEE, IAPR, OSA and AAAS. He holds three patents.  Prof. Chellappa served as the Editor-in-Chief of IEEE Transactions on Pattern Analysis and Machine Intelligence.  He has served as a General and Technical Program Chair for several IEEE international and national conferences and  workshops. He is a Golden Core Member of IEEE Computer Society and served a two-year term as a Distinguished Lecturer of IEEE Signal Processing Society. Recently, he completed a
two-year term as the President of IEEE Biometrics Council.

Leave a Reply