ScAI has a strong presence of researchers working in Computer Vision. The research focuses on several areas, including trustworthy AI, learning with
limited annotated data, autonomous driving, medical imagery, and satellite vision.
Trustworthy AI
Our goal is to make modern computer vision systems using machine learning techniques more trustworthy. Specific topics of interest include Adversarial attack and defense techniques, Reliability, and Fairness.
Learning with limited
annotated data
Our goal is to develop self-learning mechanisms that can scale the research and development of generic vision systems. We work on a variety of problems, including self-supervised learning, few-shot learning, domain adaptation, and domain generalization.
Mobility
Our goal is to develop computer vision techniques for driving assistance and autonomous driving systems in the Indian context. The topics of interest include Active Learning, Domain Adaptation, and Domain Generalization.
Egocentric vision
We are interested in novel computer vision applications arising out of body worn egocentric cameras. We are specifically interested in developing techniques to understand the wearer attributes, and his/her social life from daylong egocentric videos.
Social impact
Our vision researchers work in several applications with social impact, e.g., medical and satellite image analysis.
Faculty Members Currently Involved
