Xiaoyan Xing (邒晓岩)

I am a second-year ELLIS PhD candidate at Delta Lab, a lab funded by Bosch and University of Amsterdam, where I am working on computer vision and machine learning, and am supervised by Prof. Dr. Theo Gevers. I am also very glad to work with Dr. Jan Hendrik Metzen and Dr. Konrad Groh.

Prior to that, I obtained my master's degree (MSc) from Tsinghua University and my bachelor's degree (BEng) from Chongqing University, both in Biomedical Engineering.

I am a photography enthusiast πŸ“Έ, I play the drum in a rock band πŸ₯, and I also like traveling (20+ countries) πŸ—ΊοΈ.

I serve as a regular reviewer for CVPR, ECCV, NeurIPS, and IJCV.

Email  /  Linkedin  /  Google Scholar  /  Github

profile photo

[August 2023] One paper was accepted by CV4Metaverse (ICCV 2023 workshop)! Code released!

[April 2023] I will participate ICVSS this summer!

[March 2023] I became a reviewer of IJCV!

[February 2023] I was nominated as an ELLIS PhD!

[September 2022] I started my PhD adventure 🤞!

[August 2022] I arrived in the Netherlands!

[June 2022] I graduated with a Master's degree (MSc) from Tsinghua University.

[April 2022] I will start my Ph.D. journey in computer vision at Delta Lab (funded by Bosch), University of Amsterdam, supervised by prof. dr. Theo Gevers.

[March 2022] One paper was accepted by CVPR 2022!

[March 2022] One paper was accepted by ICPR 2022! (early accepted)


Photography is known as the art of illumination and color, in which I have great interest in. My research focus on understanding the intrinsic and invariant information from images.

Intrinsic Appearance Decomposition Using Point Cloud Representation
Xiaoyan Xing, Konrad Groh, Sezer Karaoglu, Theo Gevers
CV4Metaverse Workshop, ICCV 2023 (Oral)
arxiv / workshop (4 pages) / code

Point cloud based intrinsic appearance decomposition achieves good accuracy and can be generalized to unseen scenes/shapes.

Point Cloud Color Constancy
Xiaoyan Xing, Yanlin Qian, Sibo Feng, Yuhan Dong, Jiri Matas
CVPR, 2022
paper / code / data

Point cloud is an ideal form that contains color and depth information and can be used to improve illumination estimation pipline.

Dual-Illumination Weighting and Estimation
Xiaoyan Xing, Sibo Feng, Yanlin Qian, Yuhan Dong
ICPR, 2022

Efficient RAW dual-Illumination image generation using a random gaussian map and guided filter helps the network to estimate illumination map of unseen dual-illuminant images.

tsinghua Teaching asistant, Big Data Experiment and Application Spring 2021
Teaching asistant, Public Health Statistics Spring 2020

This website is based on Jon's source code.