Xiao Li

Beijing, China · pable-remove-etoli at gmail dot com

My name is Xiao Li (李潇). I am currently a Graphics Research Scientist at Y-tech, Kuaishou Technology (Kwai Inc.).
I got my Ph.D. degree from joint Ph.D. program between University of Science and Technology of China (USTC) and Microsoft Research Asia (MSRA) in 2019. Prior to that, I obtained my B.Eng. degree also from USTC in 2013.

From 2014 to 2019 I have a wonderful time as a research intern in Internet Graphics group at MSRA where I conducted my research on Ph.D. thesis with the topic of appearance and geometry modeling with deep learning approach.
(Here is an article about my greatest internship experience.)

Currently my research interests fall in the field of computer graphics and computer vision.
Specifically, I am interested in appearance modeling / rendering, 3D shape synthesis and generation.

Latest Resume


Publications

Deep Inverse Rendering for High-resolution SVBRDF Estimation from an Arbitrary Number of Images

SIGGRAPH 2019

Duan Gao*, Xiao Li*, Yue Dong, Pieter Peers, Kun Xu, Xin Tong (*equal contribution)
"A unified deep inverse rendering framework that enables high-resolution appearance estimation from an arbitrary number of inputs."
July 2019

Synthesizing 3D Shapes from Silhouette Image Collections using Multi-projection Generative Adversarial Networks

CVPR 2019

Xiao Li, Yue Dong, Pieter Peers, Xin Tong
"A novel multi-projection GAN framework that learns generator of 3D voxel shapes from purely unannotated 2D image collections."
June 2019

Capturing Piecewise SVBRDFs with Content Aware Lighting

CGI 2019 (Short paper track)

Xiao Li, Peiran Ren, Yue Dong, Gang Hua, Xin Tong, Baining Guo
"An adaptive active lighting design method that captures piecewise materials with two images."
April 2019

Mimicking the In-Camera Color Pipeline for Camera-Aware Object Compositing

arxiv preprint

Jun Gao, Xiao Li, Liwei Wang, Sanja Fidler, Stephen Lin
"A method for compositing virtual objects into a photograph such that the object colors appear to have been processed by the photo's camera imaging pipeline."
March 2019

Single Photograph Surface Appearance Modeling with Self-augmented CNNs and Inexact Supervision

Pacific Graphics 2018

Wenjie Ye, Xiao Li, Yue Dong, Pieter Peers, Xin Tong
"A deep-learning method that estimate material properties from a single photograph with only unlabeled training data."
July 2018

Modeling Surface Appearance from a Single Photograph using Self-Augmented Convolutional Neural Networks

SIGGRAPH 2017

Xiao Li, Yue Dong, Pieter Peers, Xin Tong
"A novel self-augment training method using few labeled images and large number of unlabeled images for appearance modeling from a single image."
July 2017

Hair Interpolation for Portrait Morphing

Pacific Graphics 2013

Yanlin Weng, Lvdi Wang, Xiao Li, Menglei Chai, Kun Zhou
"A hair morphing method based on a many-to-many strand correspondence."
July 2013

Experience

Research Scientist

Y-tech, Kuaishou Technology (Kwai Inc.)

August 2019 - Present

Research Intern

Internet Graphics group, Microsoft Research Asia

July 2014 - June 2019

Research Intern

Internet Graphics group, Microsoft Research Asia

July 2012 - June 2013

SDE Intern

Academy of Opto-Electronics, Chinese Academy of Sciences

July 2011 - Oct 2011

Education

University of Science and Technology of China

Ph.D.
Pattern Recognition & Intelligence Systems
2013 - 2019

University of Science and Technology of China

B.Eng.
Electronics Engineering
2009 - 2013

High School of Peking University

2006 - 2009
Nifty tech tag lists from Wouter Beeftink