Vijayakumar Bhagavatula
U.A. and Helen Whitaker Professor, Electrical and Computer Engineering
Affiliated Faculty, DSSC
Bio
Vijayakumar (Kumar) Bhagavatula is the U.A. and Helen Whitaker Professor of Electrical and Computer Engineering. He received a B.Tech. degree in electrical engineering from the Indian Institute of Technology, Kanpur in 1975, an M.Tech. degree in electrical engineering, also from the Indian Institute of Technology, Kanpur in 1977, and a Ph.D. in electrical engineering from Carnegie Mellon University in 1980. He joined the faculty of Electrical and Computer Engineering in 1982. He has served as acting department head of Electrical and Computer Engineering, as interim dean for the College of Engineering, and as the Director of CMU-Africa.
Education
Ph.D., 1980
Electrical Engineering
Carnegie Mellon University
MTech, 1977
Electrical Engineering
Indian Institute of Technology, Kanpur
BTech, 1975
Electrical Engineering
Indian Institute of Technology, Kanpur
Research
Machine Learning for Biometrics
One way of matching live biometric images (e.g., face images, iris images) with stored templates is to perform the cross-correlation of the two images. Professor Kumar and his students have developed several spatial frequency-domain methods to perform these cross-correlations in the presence of significant appearance variability (e.g., due to illumination changes, expression differences, etc.) in the biometric signatures.
Computer Vision for Connected Vehicles
Computer vision is a critical component for detecting objects in autonomous-driving applications. While advanced machine learning algorithms (including deep learning approaches) perform well in detecting larger objects such as cars and trucks, their performance degrades when trying to detect vulnerable road users (VRUs) such as pedestrians and bicyclists. Prof. Kumar’s research is developing methods to improve VRU detection performance through collaborative perception where multiple sensing agents (in vehicles, in roadside units, in traffic lights) can communicate information with each other to detect objects in challenging conditions such as occlusions and poor visibility.
Keywords
- Biometrics
- Computer vision
- Error correction coding
- Machine learning
- Deep learning