Skip to content
Profile Photo

Haocheng Yuan (袁浩程)

PhD Student in Computer Graphics
The University of Edinburgh

📧 hcyuan3@gmail.com
🔗 GitHub | Google Scholar

About Me

I am a third-year PhD candidate at the University of Edinburgh, supervised by Dr. Changjian Li. My research interests center on Computer-Aided Design (CAD), specifically developing computational algorithms that automatically perceive, create, and manipulate CAD content. My goal is to build human-in-the-loop systems that minimize the friction between a designer’s intent and its digital execution.

Outside of the lab, I enjoy hiking, badminton, skiing and photographing. I also play competitive games, reaching the top 100 in EUW in TFT.

Publications

DancingBox
DancingBox: A Lightweight MoCap System for Character Animation from Physical Proxies
CHI 2026🏅 Best Paper Honorable Mention
Haocheng Yuan, Adrien Bousseau, Hao Pan, Lei Zhong, and Changjian Li
An AI-powered easy MoCap with merely a phone.
DiffCSG
DiffCSG: Differentiable CSG via Rasterization
SIGGRAPH Asia 2024
Haocheng Yuan, Adrien Bousseau, Hao Pan, Chengquan Zhang, Niloy J. Mitra, and Changjian Li
Boolean operations lack derivatives, but their renderings do not.
CADTalk
CADTalk: An Algorithm and Benchmark for Semantic Commenting of CAD Programs
CVPR 2024⭐ Highlight Paper (Top 10%)
Haocheng Yuan, Jing Xu, Hao Pan, Adrien Bousseau, Niloy Mitra, and Changjian Li
Labelling non-textured shape programs through generative rendering.
Keypoint
Unsupervised Learning of 3D Semantic Keypoints with Mutual Reconstruction
ECCV 2022
Haocheng Yuan, Chen Zhao, Shichao Fan, Jiaxi Jiang, Jiaqi Yang
An unsupervised perspective for point cloud keypoints extraction.

Released under the MIT License.