I work in the broad area of multi-parametric image analysis which aims at integrating several channels of information (MRI, genetic and clinical scores) to solve a clinical or biological problem. I currently focus on three manifestations of this broad research problem:
- statistics of diffusion tensor imaging (DTI), where conventional approaches, including linear statistics, are not valid since the underlying imaging data lies on nonlinear manifolds
- combining different MR protocols to enhance tissue information to better characterize abnormality in normal appearing tissue affected by pathology; and
- automated facial expression analysis, where underlying non-linear variation in expression if well characterized and quantified, can explain differences between patients with neuro-psychiatric disorders and controls.
Face Detection and Tracking (INRIA, Rhone-Alpes)
Group theoretical methods in computer vision (Phd thesis work at IIT, Delhi, India)