|
I am a third-year PhD student at Zhejiang University, supervised by Prof. Yi Yang (IEEE Fellow, 90,000+ citations). My research focuses on Computer Vision and Machine Learning, with an emphasis on 3D Vision, Multimodal LLMs, and AIGC. My open-source projects have attracted over 2,500 GitHub stars. My academic excellence has been recognized through several honors, including the CHU KOCHEN Scholarship (Zhejiang University's highest honor), three China National Scholarships, and the First Prize in the China Graduate AI Competition (Ranked 4/2,228). I have consistently published at top-tier venues such as NeurIPS and CVPR, where my research was selected as a Highlight (top 3%). Currently, I am a Visiting Research Scholar at the Broad Institute of Harvard and MIT and the University of Chicago, collaborating with Prof. Benjamin Neale (200,000+ citations) and Prof. Siwei Chen. I am dedicated to developing scalable ML frameworks for high-dimensional multimodal data, aiming to bridge cutting-edge AI research with complex real-world applications. Email | CV | GitHub | Google Scholar | LinkedIn |
|
|
|
|
|
|
|
|
 |
Jiyuan Hu, Zechuan Zhang, Zongxin Yang, Yi Yang arXiv, 2026 paper | project page | code We introduce TRACE, a mesh-guided 3DGS editing framework for high-fidelity scene transformation with tangible reconstruction and geometry-aligned contextual video masking. |
 |
Zechuan Zhang, Zhenyuan Chen, Zongxin Yang, Yi Yang CVPR, 2026 paper We introduce IF-Edit, a tuning-free framework that repurposes pretrained image-to-video diffusion models for instruction-driven image editing. |
  |
Zechuan Zhang, Ji Xie, Yu Lu, Zongxin Yang, Yi Yang Conference on Neural Information Processing Systems (NeurIPS), 2025 paper | arXiv | project page | code | Semantic Scholar | 
We introduce ICEdit, an efficient and effective framework for instruction-based image editing. With only 1% trainable parameters (200M) and 0.1% training data (50k) compared to previous methods, ICEdit demonstrates strong generalization capabilities and is capable of handling diverse editing tasks.
Compared with commercial models such as Gemini, GPT4o, we are more open-source, with lower costs, faster speed (it takes about 9 seconds to process one image), and powerful performance.
|
  |
Ruisi Zhao, Zechuan Zhang, Zongxin Yang, Yi Yang ArXiv, 2025 paper | project page We introduce OMG3D, a novel framework that combines precise geometric control with the generative power of diffusion models, thus achieving significant improvements in visual performance. |
  |
Zechuan Zhang, Zongxin Yang, Yi Yang, Conference on Computer Vision and Pattern Recognition (CVPR), 2024 (Highlight, top 12%) paper | project page | poster | code | media (AK) | [新智元]| 
With just a single image, SIFU is capable of reconstructing a high-quality 3D clothed human model, making it well-suited for practical applications such as scene creation and 3D printing.
|
  |
Zechuan Zhang, Li Sun, Zongxin Yang, Ling Chen, Yi Yang Conference on Neural Information Processing Systems (NeurIPS), 2023 paper | project page | code | media (CVer)| 
A transformer-based implicit function for single image clothed human reconstruction.
|
|
|
|
|
|
|
|
template adapted from this awesome website |