About
I work on 3D computer vision, building AI systems capable of interacting with real-world 3D environments using visual inputs, much like how humans do. This includes inferring shape, materials, cameras, lighting, motion, and functional properties from limited 2D observations.
I am currently a PhD student at HKU, advised by Shenghua Gao and working closely with Yi Ma. From 2020 to 2024, I was a PhD student at ShanghaiTech University. From Winter 2022 to Spring 2023, I interned at MSRA, working with Xin Tong and Jiaolong Yang. I did my undergraduate at SCUT in 2020.
Research
See Google Scholar for a complete list of publications.
* indicates equal contribution.
|
CUPID: Pose-Grounded Generative 3D Reconstruction from a Single Image
Preprint, 2025
Create canonically posed 3D object and an object-centric camera from single image in just a few seconds. |
|
GenFusion: Closing the Loop between Reconstruction and Generation via Videos
CVPR, 2025
Generative scene inpainting with a reconstruction-driven video generation model. |
|
2D Gaussian Splatting for Geometrically Accurate Radiance Fields
SIGGRAPH, 2024 (Most Influential Paper#1)
Using surfels and a ray-cast based differentiable rasterizer enables efficient and high-fidelity geometry and apperance reconstruction from real images. |
|
Mip-Splatting: Alias-free 3D Gaussian Splatting
CVPR, 2024 (Oral, Best Student Paper)
Mip-filters enables synthesizing alias-free scenes with 3D Gaussian Splatting. |
|
3D-aware Image Generation using 2D Diffusion Models
ICCV, 2023
Create 3D consistent scene from a single image via iterative multiview RGBD sampling and fusion. |
Service
Journal reviewer: TPAMI, TVCG ...
Conference reviewer: SIGGRAPH, SIGGRAPH Asia, CVPR, ICCV, ECCV, ICLR, NeurIPS ...