Talk Abstract: Looking at just a single image of a scene is sufficient for us to build rich mental representations, enabling us to reason about the underlying 3D world. Inferring such representations from little information is a core aspect of intelligence. It helps us understand and interact with our surroundings and each other. My research aims to build similar computational models—methods that can perceive the rich 3D structured world from sparse observations such as images and videos. I will discuss the main challenges in 3D perception and how my research addresses them by posing vision as an inverse graphics problem. I will demonstrate results on perceiving humans, objects, and large unconstrained scenes.

Speaker Biography: Ayush Tewari is a postdoctoral researcher at MIT CSAIL with Bill Freeman, Vincent Sitzmann, and Josh Tenenbaum. He previously completed his Ph.D. at the Max Planck Institute for Informatics, advised by Christian Theobalt. His research interests lie at the intersection of computer vision, machine learning, and computer graphics, focused on 3D perception. He was awarded the Otto Hahn medal from the Max Planck Society for his scientific contributions as a Ph.D. student.