Brief Intro.

Zheng Dong is currently a research fellow (特聘研究员) at Zhejiang Sci-Tech University. He received his Ph.D. Degree in computer vision and graphics from the State Key Laboratory of CAD&CG, Zhejiang University, under the supervision of Prof. Weiwei Xu (Zhejiang University) and Prof. Rynson W.H. Lau (City University of Hong Kong). His research interests lie in human-centric 3D/4D vision and 2D image processing, with a focus on image-based 3D human reconstruction, dynamic human-scene modeling, and reflection/flare removal in 2D images, among others.

E-mail: zhengdong AT zstu.edu.cn / zju.edu.cn

WeChat: -zhengdong-

Resume: [resume.pdf]

Academic Service

Reviewer: SIGGRAPH, CVPR, ICCV, NeurIPS, IJCV, TVCG, CAVW, etc

Background

Researcher | Apr. 2025 - Present
School of Artificial Intelligence
Zhejiang Sci-Tech University (浙江理工大学)


Ph.D. Degree in Computer Vision and Graphics | Sep. 2019 - Mar. 2025
State Key Laboratory of CAD&CG
Zhejiang University (浙江大学)


Bachelor of Computer Science | Sep. 2015 - Jun. 2019
School of Computer Science and Engineering
Central South University (中南大学)

Publications

Gaussian Surfel Splatting for Live Human Performance Capture
Zheng Dong, Ke Xu, Yaoan Gao, Hujun Bao, Weiwei Xu, Rynson W.H. Lau
ACM Trans. on Graphics, 43(6), (Proc. ACM SIGGRAPH Asia 2024)
Project / Paper / Video / Github
We propose a novel regressed-surface-point-based method, to leverage Gaussian surfel splatting and blending-based appearance enhancement, for human live free-view rendering from 4 RGBD sensors.


SAILOR: Synergizing Radiance and Occupancy Fields for Live Human Performance Capture
Zheng Dong, Ke Xu, Yaoan Gao, Qilin Sun, Hujun Bao, Weiwei Xu, Rynson W.H. Lau
ACM Trans. on Graphics, 42(6), (Proc. ACM SIGGRAPH Asia 2023)
Project / Paper / Video / Github
SAILOR is a generalizable method for human free-view rendering and reconstruction from very sparse (e.g., 4) RGBD streams, achieving near real-time performance under acceleration.


Geometry-aware Two-scale PIFu Representation for Human Reconstruction
Zheng Dong, Ke Xu, Ziheng Duan, Hujun Bao, Weiwei Xu, Rynson W.H. Lau
Annual Conference on Neural Information Processing Systems (NeurIPS) 2022 (Spotlight)
Project / Paper / Video / OpenReview
The geometry-aware two-scale PIFu is a method to address the problem of PIFu-based approaches that reconstruct flat facial and geometry-fallible bodies in a sparse capture setting.


Location-aware Single Image Reflection Removal
Zheng Dong, Ke Xu, Yin Yang, Hujun Bao, Weiwei Xu, Rynson W.H. Lau
International Conference on Computer Vision (ICCV) 2021
Paper / Supp. / Video / Github
LARR leverages the Laplacian features to regress probabilistic reflection confidence maps, and incorporates a recurrent network to refine reflection removal results at each iteration.

Distinctions

Outstanding Graduate Student,Zhejiang University (2025),浙江大学优秀毕业生

Excellent Postgraduate Student Award, Zhejiang University (2020, 2022-2023),浙江大学优秀研究生

Outstanding Graduate Student, Central South University (2019),中南大学优秀毕业生

National Scholarship, Central South University (2016),国家奖学金