PhD Student · HKU

Yuan Ze (袁泽)

Generating things, from 3D textures to videos.

I am a PhD student at The University of Hong Kong, advised by Prof. Xiaojuan Qi.

I graduated from Beihang University (M.S.) in Jan 2025, supervised by Prof. Ke Xu.

I have always been interested in generating colorful things. My previous work focused on 3D texture generation, and I am now exploring video generation.

Email Google Scholar GitHub ORCID

Research

Current direction

Video Generation

I am currently exploring video generation and how generative models can produce coherent, vivid, and controllable dynamic content.

Previous work

3D Texture Generation

My previous research generated high-quality and consistent textures for 3D assets by building on the visual priors of image and video diffusion models.

The common thread is simple: I enjoy generating colorful things. Moving from 3D textures to videos feels like a natural continuation of that interest.

Publications

Full list on Google Scholar

A gallery of colorful textured 3D assets generated by SeqTex

First Author

SeqTex: Generate Mesh Textures in Video Sequence

Ze Yuan, Xin Yu, Yangtian Sun, Yuan-Chen Guo, Yan-Pei Cao, Ding Liang, Xiaojuan Qi

SIGGRAPH Asia 2025 · Conference Track · 2025

SeqTex formulates mesh texture generation as a sequence generation problem and leverages video diffusion priors for coherent texture synthesis.

Paper Project DOI

A richly textured bird generated with TEXGen

Second Author

TEXGen: a Generative Diffusion Model for Mesh Textures

Xin Yu, Ze Yuan, Yuan-Chen Guo, Ying-Tian Liu, Jianhui Liu, Yangguang Li, Yan-Pei Cao, Ding Liang, Xiaojuan Qi

ACM Transactions on Graphics · SIGGRAPH Asia 2024 · Journal Track · 2024 (Best Paper Honorable Mention)

TEXGen is a generative diffusion model for high-quality mesh textures that builds on pretrained image diffusion priors.

Paper Project DOI

First Author

VRDistill: Vote Refinement Distillation for Efficient Indoor 3D Object Detection

Ze Yuan, Jinyang Guo, Dakai An, Junran Wu, He Zhu, Jianhao Li, Xueyuan Chen, Ke Xu, Jiaheng Liu

ACM Multimedia 2024 · Conference Paper · 2024

VRDistill introduces vote refinement distillation for efficient indoor 3D object detection.

Paper

Background

Education

Experience

education

2025.06–Present

Ph.D. Student · The University of Hong Kong

Advised by Prof. Xiaojuan Qi.

experience

2023–Present

Research Intern · Vast AI

AI research and software engineering.

education

2022–2025.01

M.S. · Beihang University

Advised by Prof. Ke Xu.

experience

2020–2021

Software Engineer · TP-Link

Java web development.

education

2016–2020

B.S. · Dalian University of Technology

I hold JLPT N2 and enjoy Japanese dramas, films, and anime.

Contact

For research conversations and collaborations, email is the best way to reach me.

Email: [email protected]
WeChat: yz-_-1998
Location: Hong Kong SAR