My research interests lie in the fields of Computer Vision, Multimedia, Machine Learning, and Natural Language Processing. I have worked on Human Centric Understanding (including hand gesture recognition, hand pose estimation, human pose estimation and tracking, human ReID, action detection, etc.), Multi-modal Joint Understanding (including image captioning, video captioning, visual-semantic embedding, etc.), shape understanding, adversarial machine learning, etc. My current focuses are: 1. human centric understanding (pose, hand, gesture, human Re-ID, and tracking); 2. object detection, action detection and video representation learning, 3. multi-modal joint understanding, vision and language.