Xiu Li  

(pronouced as "Hsiu Lee")

Ph.D, EECS
Researcher, Bytedance
Email: lixiulive [at] gmail [dot] com

Google Scholar / Github

Photo

My research interests lie broadly in computer vision,computer graphics and machine learning. I'm focusing on 3D reconstruction and synthesis, particularly on neural rendering and digital human avatar. The goal of my research is to develop methods for faithfully capturing, manipulating and synthesizing our real world.

Research


WarpDiffusion
WarpDiffusion: Efficient Diffusion Model for High-Fidelity Virtual Try-on
uabc
Neural Capture of Animatable 3D Human from Monocular Video
NReF
Estimating Neural Reflectance Field from Neural Radiance Field using Tree Structures
uabc
Universal and Flexible Optical Aberration Correction Using Deep-Prior Based Deconvolution
mmmcap
4D Association Graph for Realtime Multi-person Motion Capture Using Multiple Video Cameras
densecap
Full-body Markless Motion Capture with Full-body Parsing
sfrm
Structure from Recurrent Motion:From Rigidity to Recurrency

Short Bio (cv)

    Dr. Xiu Li is a Researcher at Bytedance. He received Ph.D. degree from Tsinghua University in 2022. He was a visiting scholar in School of Computer Science at Carnegie Mellon University from Sep. 2017 to Aug, 2019, working with Prof.Yaser Sheikh and Prof. Hongdong Li. Before that, Xiu received B.E. in Department of Automation,Tsinghua University. He has also been with Microsoft Research Asia and Tencent. His research spans from 3D Vision to computational photography and machine learning. He was with the champion team of 2013 International Aerial Robotics Competition.