About Me

I am a second-year master’s student at Tsinghua University, advised by Prof. Xiu Li. I also conduct research with Vision and Image Processing (VIP) Laboratory, Duke University, advised by Prof. Sina Farsiu. Before that, I received my B.Eng. degree in Software Engineering from Southwest University in 2024, where I graduated with the “Special Scholarship” and was named an “Outstanding Student Representative.” My research interests revolve around the intersection of vision restoration and perception, visual understanding, and multi-agent system.

I am currently seeking PhD positions for Fall 2027 and open to all forms of collaboration.

Education

Tsinghua University
Master of Engineering in Artificial Intelligence
Southwest University
Bachelor of Engineering in Software Engineering

Experiences

Global Monetization GenAI, TikTok
Research Scientist Intern
Duke University
Research Assistant
DAMO Academy, Alibaba
Research Scientist Intern

Selected Publications

* Equal contribution, † Corresponding author. You can find more paper in my Google Scholar.

ICLR 2026
sym

Photon: Speedup Volume Understanding with Efficient Multimodal Large Language Models

Chengyu Fang* , Heng Guo*, Zheng Jiang, Chunming He, Xiu Li†, Minfeng Xu†

  • Photon is a variable-length 3D medical VQA framework with instruction-conditioned token scheduling and surrogate gradients, achieving adaptive acceleration and state-of-the-art performance.
NeurIPS 2024
Spotlight
sym

Real-world Image Dehazing with Coherence-based Pseudo Labeling and Cooperative Unfolding Network

Chengyu Fang* , Chunming He*†, Fengyang Xiao, Yulun Zhang†, Longxiang Tang, Yuelin Zhang, Kai Li, and Xiu Li†

  • The cooperative unfolding network (CORUN) and the first plug-in-play iterative mean-teacher framework (Colabator) for real-world image dehazing.
  • Selected as VALSE 2025 Top 10 Popular Poster
ICLR 2025
Spotlight
sym

Reti-Diff: Illumination Degradation Image Restoration with Retinex-based Latent Diffusion Model

Chunming He*, Chengyu Fang*†, Yulun Zhang†, Kai Li, Longxiang Tang, Chengyu You, Fengyang Xiao, Zhenhua Guo, Xiu Li and Sina Farsiu†

  • The first latent diffusion model-based methods with strong generalizability in illumination degradation image restoration problems and promising performance in downstream tasks.
ArXiv 2026
sym

PRISM: Rethinking Scattered Atmosphere Reconstruction as a Unified Understanding and Generation Model for Real-world Dehazing

Chengyu Fang, Chunming He, Yuelin Zhang, Chubin Chen, Chenyang Zhu, Hongqiu Wang, Longxiang Tang, Xiu Li†, and Sina Farsiu†

  • PRISM is a real-world dehazing framework that jointly reconstructs clear scenes and scattering variables, while bridging the sim2real domain gap through selective self-distillation and self-reinforcing prior.
ICLR 2026
sym

Annotation-Free Medical Visual Reasoning via Agentic Reinforcement Learning

Zheng Jiang*, Heng Guo*, Chengyu Fang* , Changchen Xiao, Xinyang Hu, Lifeng Sun†, Minfeng Xu†

  • MedVR is the first end-to-end reinforcement learning framework that integrates visual and textual reasoning for medical VLMs, eliminating the need for costly intermediate supervision.
TPAMI 2025
sym

Diffusion Models in Low-Level Vision: A Survey

Chunming He*†, Yuqi Shen*, Chengyu Fang* , Fengyang Xiao, Longxiang Tang, Yulun Zhang, Wangmeng Zuo, Zhenhua Guo, Xiu Li†

  • A curated list of awesome Diffusion Models(DMs) in low-level vision.
arXiv 2025
sym

MultiCOS: Unlocking the Potential of Limited Multimodal Data in Camouflaged Object Segmentation

Chengyu Fang* , Chunming He*, Yuqi Shen, Chenyang Zhu, Yuelin Zhang, Fengyang Xiao, Longxiang Tang, Chubin Chen, Xiu Li†

  • A novel framework that effectively leverages diverse data modalities to improve segmentation performance.
arXiv 2025
sym

M3Ret: Unleashing Zero-shot Multimodal Medical Image Retrieval via Self-Supervision

Che Liu*, Zheng Jiang*, Chengyu Fang* , Heng Guo, Yan-Jie Zhou, Jiaqi Qu, Le Lu, Minfeng Xu†

  • A unified visual encoder without any modality-specific customization for various medical visual modalities in 2D and 3D.

Teaching

  • 2023.09 - 2024.01 & 2024.09 - 2025.01  Teaching Assistant for Frontiers of AI technology and industrial applications, Tsinghua University.

Honors and Awards

  • 2026.05  ICML 2026 Silver Reviewer
  • 2024.06  Chongqing Outstanding Graduates
  • 2024.05  Representative of National Scholarship (Published in China’s official media People’s Daily).
  • 2024.04  Chongqing Merit Student
  • 2023.12  Southwest University Outstanding Student Representative
  • 2023.04  Chongqing Advanced Individual for Innovation Capability

Scholarships

  • 2023.12  National Scholarship
  • 2022.12  National Scholarship
  • 2023.12  Xiaomi Corporation Special Scholarship
  • 2023.12  Southwest University Special Scholarship
  • 2022.07  Professor Qiu Yuhui Scholarship
  • 2022.07  Pisen Electronics Co. Ltd Scholarship
  • 2021.10  Southwest University First Class Scholarship

Competition

  • 2023.08  🏅1st Prize of “Texas Instruments Cup” 2023 National Undergraduate Electronic Design Contest
  • 2023.08  🏅1st Prize of “China Software Cup” University Student Software Design Competition
  • 2023.08  🏅1st Prize of “China University Student Embedded Chip and System Design Competition
  • 2023.04  🏅1st Prize of 2023 China University Robot Competition (RoboMaster RMUL)
  • 2022.08  🏅️1st Prize of “China Software Cup” University Student Software Design Competition
  • 2022.12  🏅1st Prize of 2022 China University Robot Competition (RoboMaster RMUL)
  • 2023.06  🥈2nd Prize in China Robotics and Artificial Intelligence Competition
  • 2022.08  🥈2nd Prize of “China Software Cup” University Student Software Design Competition
  • 2022.06  🥈2nd Prize of 2022 China University Robot Competition (RoboMaster RMUT)
  • 2023.08  🥉3rd Prize in Chinese Collegiate Computing Competition

Invited Talk

  • 2026.04  “Speedup Volume Understanding with Efficient Multimodal Large Language Models”, AI TIME

My Friends, Collaborators, and Long-term Cooperative Professors