Rama Chellappa


Primary Appointment: Department of Electrical and Computer Engineering and Department of Biomedical Engineering

Research Interests

  • Computer vision
  • Pattern recognition
  • Machine learning
  • Artificial intelligence
  • Signal and image processing

My research interests span computer vision, pattern recognition, machine learning and artificial intelligence. We integrate concepts from 3D geometry, illumination models, sensor physics, differential geometry, knowledge representation and reasoning methods, sparse and deep representations for addressing problems in these areas. In the area of computer vision, my group has worked on Markov random fields and extensions, recovering 3D structure from a single image, and from long sequences using discrete and continuous approaches, hyperspectral and radar image analysis. We are also interested in computer vision approaches to medicine. Specifically, we are interested in marker less motion capture methods and gait analysis with applications in diagnosing movement-related disorders. We have also developed registration methods that help electrophysiologists perform efficient ablation procedures. We are currently working on geolocation of images, media forensics and medical image reconstruction.

In the area of pattern recognition and machine learning, my group is working on many problems related to object/face/action detection and recognition from still images and videos. We use novel approaches based on shallow and deep representations to build robust end-to-end systems for face, expression, gesture, object and action detection and recognition. We are currently working on many current topics such as unsupervised domain adaptation, domain generalization, multi-task learning, learning from few samples, Generative Adversarial Networks (GANs) and defending attacks on machine learning algorithms and systems. We are also working on bias mitigation approaches for designing fair machine learning systems.

In the area of artificial intelligence, we have worked on topics such as knowledge representation, random forests, stochastic petri nets, non-monotonic reasoning and ontologies with potential applications in computer vision and medicine. We are currently working on the “deepfakes” problem and knowledge representation in deep learning networks.

Center for Language and Speech Processing