We develop intelligent algorithms for automated interpretation of images and video. Our research ranges from low-level processing to high-level semantic interpretation. We work on diverse topics such pose estimation, activity recognition and segmentation.