Ph.D. researcher · HUST

Jingfeng Yao 姚劲枫

I build generative models and the visual representations that make them work.

I am a third-year Ph.D. student at Huazhong University of Science and Technology, supervised by Prof. Xinggang Wang, and expected to graduate in 2027. My work spans generative models, image matting, and representation learning.

Email me GitHub Google Scholar

First-author papers: 10
GitHub stars: —+
My ViTMatte model downloads: 50M+

Portrait of Jingfeng Yao — Wuhan, China Class of 2027

Research

Selected publications

Work across visual generation, efficient learning, and representation.

Most Influential CVPR Paper · 11th

Generative models CVPR 2025

Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models

Jingfeng Yao, Bin Yang, Xinggang Wang

CVPR 2025 Oral. Accepted with full score.

Ranked 11th among the most influential CVPR 2025 papers by Paper Digest.

arXiv Code

My model · 50M+ downloads

Image matting Information Fusion 2023

ViTMatte: Boosting Image Matting with Pretrained Plain Vision Transformers

Jingfeng Yao, Xinggang Wang, Shusheng Yang, Baoyuan Wang

Integrated into Hugging Face Transformers as a standard matting method.

The ViTMatte model I trained was integrated into Nuke, and its weights have been downloaded over 50 million times.

arXiv Code

More first-author work

* Equal contribution

Visual overview of Towards Scalable Pre-training of Visual Tokenizers

Visual tokenizersECCV 2026

Towards Scalable Pre-training of Visual Tokenizers for Generation

Jingfeng Yao, Yuda Song, Yucong Zhou, Xinggang Wang

arXiv Code

Vision-language modelsECCV 2026

DiffusionVL: Translating Any Autoregressive Models into Diffusion Vision Language Models

Lunbin Zeng*, Jingfeng Yao*, Bencheng Liao, Hongyuan Tao, Wenyu Liu, Xinggang Wang

arXiv Code

Medical AIarXiv 2026

UniX: Unifying Autoregression and Diffusion for Chest X-Ray Understanding and Generation

Ruiheng Zhang*, Jingfeng Yao*, Huangxuan Zhao*, Hao Yan, Xiao He, Lei Chen, Zhou Wei, Yong Luo, Zengmao Wang, Lefei Zhang, Dacheng Tao, Bo Du

arXiv Code

Efficient video generationAAAI 2026

Turbo-VAED: Fast and Stable Transfer of Video-VAEs to Mobile Devices

Ya Zou*, Jingfeng Yao*, Siyuan Yu, Shuai Zhang, Wenyu Liu, Xinggang Wang

arXiv Code

Efficient trainingNeurIPS 2024

FasterDiT: Towards Faster Diffusion Transformers Training without Architecture Modification

Jingfeng Yao, Cheng Wang, Wenyu Liu, Xinggang Wang

arXiv Code

Medical AInpj Digital Medicine

EVA-X: A Foundation Model for General Chest X-ray Analysis with Self-supervised Learning

Jingfeng Yao, Xinggang Wang, Yuehao Song, Huangxuan Zhao, Jun Ma, Yajie Chen, Wenyu Liu, Bo Wang

arXiv Code

Image mattingIVC 2024

Matte Anything! Interactive Natural Image Matting with Segment Anything Models

Jingfeng Yao, Xinggang Wang, Lang Ye, Wenyu Liu

arXiv Code

Medical AIarXiv 2024

LKCell: Efficient Cell Nuclei Instance Segmentation with Large Convolution Kernels

Ziwei Cui*, Jingfeng Yao*, Lunbin Zeng, Juan Yang, Wenyu Liu, Xinggang Wang

arXiv Code

Recognition

Awards

2024

China National Scholarship

国家奖学金

Awarded to only 0.2% of candidates nationwide.

2022

Gold Award, China College Students' “Internet+” Competition

中国大学生“互联网+”创新创业大赛国赛金奖

About 300 Gold Awards from 3.4 million projects, a 0.009% success rate.

Jingfeng Yao during his year as a volunteer teacher in Lincang

Beyond research

A year in the classroom

I spent one year as a volunteer teacher in Lincang, Yunnan Province, representing Huazhong University of Science and Technology. I taught history at Linxiang No. 1 Middle School.

Read the news coverage