He is a computer vision researcher with a primary interest in geometry. Inter alia he has introduced and developed the group-theoretic approach of motion averaging.
The current focus of his research group is efficient and accurate large-scale 3D reconstruction using both RGB and depth images, i.e. problems in structure-from-motion and scan modeling respectively. He is also interested in photometric stereo, principled approaches for image and mesh denoising, and tensor decomposition methods for higher-order grouping.