Yifan Li

Hi!😁 I’m Yifan Li (李一凡). I am currently a first-year PhD student at Peking University. I am honored to join in the Spatial and Temporal Restoration, Understanding and Compression Team (STRUCT) since 2023, supervised by Prof. Jiaying Liu, Prof. Shuai Yang and Prof. Wenhan Yang. I obtained my Bachelor’s degree in Data Science from Yuanpei College, Peking University.

My current research interests lies at the intersection of generative models and robust machine understanding against degradation, where I try my best to construct a more photorealistic, touchable, and intelligent world.

Towards a photorealistic and touchable visual expression, I focus on the language-based colorization during my undergraduate studies and achieve lifelike and highly-controllable recoloring with rich applications. (One published at ACM MM’24 and Two under review). For more natural and versatile editing on videos, I propose a self-supervised relighting framework for extendable video chroma-lux editing (One Under Review). Target on clearer and brighter photos, I propose a zero-shot low-light image enhancement method (One published at IEEE TPAMI’25) with efficient application.

Towards robust degradation-free intelligence, I observe the impact priors embeds in Vision Foundation Models and physical laws, and propose an insightful zero-shot domain adaptation scheme under low-light conditions with large-margin SOTA performance and superior practical utility (One accepted by CVPR’26).

Now, I am exploring further possibilities to unleash generative potentials on human perception-centric low-light image enhancement, and semantic-guided RL post training for Diffusion-based Super-Resolution Models.

Feel free to reach out to me at Email if you have any questions, or want to know more about me!

selected publications

arXiv 26.04

VibeFlow: Versatile Video Chroma-Lux Editing through Self-Supervised Learning

Yifan Li, Pei Cheng, Bin Fu, Shuai Yang, and Jiaying Liu

arXiv:2604.13425, 2026

Bib HTML PDF

@article{vibeflow,
  title = {{VibeFlow}: Versatile Video Chroma-Lux Editing through Self-Supervised Learning},
  author = {Li, Yifan and Cheng, Pei and Fu, Bin and Yang, Shuai and Liu, Jiaying},
  journal = {arXiv:2604.13425},
  year = {2026},
}

ACM MM

COCO-LC: Colorfulness Controllable Language-based Colorization

Yifan Li, Yuhang Bai, Shuai Yang, and Jiaying Liu

In ACM MM, 2024

Bib PDF Code

@inproceedings{coco-lc,
  title = {{COCO-LC}: Colorfulness Controllable Language-based Colorization},
  author = {Li, Yifan and Bai, Yuhang and Yang, Shuai and Liu, Jiaying},
  booktitle = {ACM MM},
  year = {2024},
}

CVPR

Towards Generalized Representations for Low-Light Understanding: When Signal Constancy Meets Semantic Enrichment

Yifan Li, Haofeng Huang, Wenhan Yang, and Jiaying Liu

In CVPR, 2026

Bib Code

@inproceedings{uniprior,
  title = {Towards Generalized Representations for Low-Light Understanding: When Signal Constancy Meets Semantic Enrichment},
  author = {Li, Yifan and Huang, Haofeng and Yang, Wenhan and Liu, Jiaying},
  booktitle = {CVPR},
  year = {2026},
}

TPAMI

QuadPrior++: Multi-dimension Augmented Physical Prior for Zero-Reference Illumination Enhancement

Haofeng Huang^*, Yifan Li^*, Wenjing Wang, Wenhan Yang, Ling-Yu Duan, and Jiaying Liu

IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2025

Bib PDF Code

@article{quadprior++,
  title = {{QuadPrior++}: Multi-dimension Augmented Physical Prior for Zero-Reference Illumination Enhancement},
  author = {Huang, Haofeng and Li, Yifan and Wang, Wenjing and Yang, Wenhan and Duan, Ling-Yu and Liu, Jiaying},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)},
  year = {2025},
}