Ramalingam Chellappa

From ETHW

Ramalingam Chellappa
Ramalingam Chellappa

Biography

A pioneer in developing image, video, and multidimensional signal processing theory and methods for solving image processing, computer vision, and pattern recognition problems, Ramalingam Chellappa has profoundly affected the development of systems for face recognition and verification, image and video synthesis and analytics, real-time action detection, and active authentication. During the 1980s, Chellappa developed groundbreaking parameter estimation and neighborhood selection rules for Markov random field (MRF) models and developed algorithms for image restoration, texture analysis, and segmentation using MRFs. His linear discriminant analysis-based algorithm developed during the 1990s pioneered discriminative methods for training face recognition systems. Chellappa has also been active in creating tools for 3D recovery from one or more images using discrete and continuous methods. The Frankot-Chellappa algorithm was developed for extracting integrable 3D surfaces from a single image. He pioneered the area of video-based 3D modeling algorithms using batch and recursive estimation methods. He was also an early contributor to neural networks for image processing and computer vision applications. His recent work in this area includes deep-learning-based face, object, and activity recognition systems and methods that combat attacks against classification systems. Specifically, he has developed deep convolution neural network (DCNN)-based algorithms to help overcome the problems of aging, pose, and illumination, affecting accurate face recognition and providing results that rivaled human performance. His work on the HyperFace and UltraFace algorithms resulted in programs that perform not only face detection and classification but also gender recognition and age and pose estimation. The end-to-end system built by his group has been used for child exploitation/crime prevention and other U.S. homeland security applications.

An IEEE Life Fellow and recipient of the 2012 K.S. Fu Prize from the International Association for Pattern Recognition, Chellappa is a Distinguished University Professor with the Electrical and Computer Engineering Department at the University of Maryland, College Park, MD, USA.