My research centers on the visual perception of people. People are arguably one of the most important types of content in images and video, and many intelligent systems rely on the perception and the interpretation of people in the real world. My research program covers both high-level semantics and lower-level physical modelling of people. On the semantics side, my research centers on video understanding relating to human actions. On the physical side, my research centers on human 3D modelling, covering topics such as 3D pose and 3D shape estimation of the body and hand.