Cholakkal’s research aims at developing solutions that are well-aligned with real-world applications and is focused on four areas within computer vision: visual recognition; person-centric scene understanding; learning from limited supervision; and image generation. His recent research interests include object detection; image and video segmentation; object counting; image classification; pedestrian detection; person search; human-pose estimation; human-object interaction detection; activity recognition; crowd counting; few-shot/zero-shot learning; weakly supervised learning; AI for style imitation and AI for creativity.