Portrait of Bowen Zhang Tencent Hunyuan · 3D
// Researcher · 3D Generative AI

Hi, I'm Bowen Zhang

Researcher at Tencent Hunyuan, building the Hunyuan3D family of 3D foundation models.

My research lies at the intersection of computer vision, computer graphics and generative AI. I focus on large-scale 3D asset generation — from compact 3D latent representations and image/text-to-3D diffusion, to production-ready PBR materials and end-to-end pipelines for game-ready assets. Before joining Tencent, I studied at Xi'an Jiaotong University with Prof. Xi Zhao, working on point cloud completion and 3D shape understanding.

Google Scholar GitHub Publications

01Research Interests

What I'm working on

3D Generative Models

Scalable native 3D generation from images and text, compact 3D latent spaces, and multi-modal controllable generation.

Diffusion3D VAEImage-to-3DText-to-3D

Production-Ready 3D Assets

High-fidelity PBR texture synthesis, end-to-end AI pipelines for game-ready assets, and benchmarks for 3D generation.

PBR MaterialTexturingPipelineBenchmark

3D Shape Understanding

Point cloud completion, primitive-based shape abstraction, and 3D human pose reconstruction from partial observations.

Point CloudCompletionPrimitivesRGB-D

02News

Recent updates

  • Released HY3D-Bench, a benchmark for 3D asset generation. [arXiv]NEW

  • Released Hunyuan3D-Omni for controllable 3D generation and Hunyuan3D Studio, an end-to-end pipeline for game-ready assets.

  • Open-sourced Hunyuan3D 2.1 with production-ready PBR materials. [Code]

  • Released Hunyuan3D 1.0, a unified framework for text-to-3D and image-to-3D generation.

  • Compress3D accepted to ECCV 2024.

  • Shape Completion with Points in the Shadow accepted to SIGGRAPH Asia 2022. [Project]

03Publications

Full list on Google Scholar
2026
2025
2024
2023
2022
2021

No publications match your search.

04Experience & Education

Researcher

Present
Tencent Hunyuan

Core member of the Hunyuan3D team, working on large-scale 3D generative models, PBR texture generation and production pipelines for 3D assets.

Graduate Study

Xi'an, China
Xi'an Jiaotong University

Advised by Prof. Xi Zhao. Research on point cloud completion, 3D shape abstraction and single-image 3D generation.

Let's build the future of 3D together.

I'm always happy to discuss research ideas, collaborations and opportunities in 3D generation, graphics and vision.

Scholar GitHub