Computer Vision Seminar Series: Dr. Chuanxia Zheng, "Beyond Visual Geometry: Toward Physical 3D."
- Sponsor
- Illinois Computer Vision
- Speaker
- Dr. Chuanxia Zheng
- Contact
- Yao Xiao
- yaox11@illinois.edu
- Originating Calendar
- Siebel School Speakers Calendar
Abstract: In this talk, I will present a series of work that go beyond visual (visible) geometry. I will argue that dealing with 3D structures is essential for understanding, interacting with, and building the physical world. I will first introduce Amodal3R and NOVA3R, which recover complete, physically plausible 3D, including occluded regions beyond what is directly visible. Next, I will discuss 3D generation for simulation and function design, where models produce stable and functionally meaningful 3D assets. Finally, I will consider the 4D reconstruction, where the goal is to recover 3D shape motion from videos, and demonstrate how foundation video diffusion model can contribute to this task.
Speaker Bio.: Chuanxia Zheng (https://physicalvision.github.io/people/~chuanxia) is a Nanyang Assistant Professor at Nanyang Technological University (NTU), where he leads the Physical Vision Group (PVG) since 2025. Before joining NTU, he is a Marie Skłodowska-Curie Actions (MSCA) fellow in Visual Geometry Group (VGG) at the University of Oxford. He is also a DAAD Ainet Fellow of Germany, a National Research Foundation (NRF) fellow of Singapore. He received his Ph.D. degree from NTU. His research focuses on Spatial AI and Physical AI, with the goal of developing systems that can perceive, reconstruct and interact with the physical world. The broader goal is to create realistic digital twins of the physical natural world, capturing diverse physical properties, such as geometry, appearance, occlusion, motion, interaction, gravity, sound and more.