Baiang Li | Princeton University

About Me

I am a PhD student at Princeton University, advised by Felix Heide. I have the honor to collaborate closely with Jian Wang @ Snap Research and Xiaogang Xu @ CUHK.

My current research mainly focuses on 2D/3D vision, image generation models, and video generation models. As side projects, I also work on computational photography and trustworthy image/video processing tasks. Recently, I am particularly curious about how constraints imposed on latent representations can facilitate models' understanding, learning, and generation across visual tasks.

We co-founded a platform for beginners to learn about low-level vision and imaging, summarizing the latest research and techniques and holding events worldwide! There are also available resources for job opportunities. If you are interested, want to join our community, or want to post your job opportunities, please check it out at lowlevelcv.com.

I am open to collaborations! Please feel free to reach out by email if you want to work with me or intern in our Computational Imaging Lab — I have lots of interesting ideas to explore.

Research Interests

Trustworthy Imaging

Rather than directly predicting RGB pixels, we reformulate imaging tasks in novel ways that process and preserve authentic RGB content — keeping results faithful to the real captured signal.

Representation Learning

Exploring how constraints on latent representations can facilitate models' understanding, learning, and generation across visual tasks.

Generative Vision

Designing image and video generation models with controllable structure, semantics, and temporal coherence for 2D and 3D visual content.

News

Apr 2026New work released on knowledge distillation. See our project page.
Apr 2026We will present our new Weak-to-Strong project poster in AIX SUMMIT.
Mar 2026Our paper Lucky High Dynamic Range Smartphone Imaging accepted to TOG (SIGGRAPH 2026 journal track). Project page: lucky-hdr.github.io.
Mar 2026We got 3rd place (among over 200 participants) in the CVPR Image Super Resolution (x4) challenge.
Sep 2025One paper about hierarchical modeling accepted to NeurIPS. Congrats to collaborators!
Jun 2025We will post our paper in RSS 2025 Workshop on Robot Planning in the Era of Foundation Models.
May 2025Our paper about image enhancement accepted to ICCP as oral presentation.
May 2025We release our new work about autonomous driving.
May 2025Our paper about image deraining accepted to TMM.
Feb 2025Officially accepted as a PhD student at Princeton University!
Jul 2024Visiting the Computational Imaging Lab @ Princeton as a research assistant.
Jun 2024New work released on low dynamic range image enhancement.
Mar 2024One paper about image matting accepted to SIGGRAPH 2024.
Feb 2024New work released on image deraining.
Nov 2023Starting internship at OpenMMLab @ Shanghai AI Lab.
Aug 2023Paper on low-light stereo image enhancement accepted to ACM-MM 2023.

Calendar

Publications

arXiv

Weak-to-Strong Knowledge Distillation Accelerates Visual Learning

Baiang Li, Wenhao Chai, Felix Heide

Paper Project Page Poster

ACM TOG (SIGGRAPH 2026)

Lucky High Dynamic Range Smartphone Imaging

Baiang Li†, Ruyu Yan†, Ethan Tseng, Zhoutong Zhang, Adam Finkelstein, Jiawen Chen, Felix Heide († Equal contribution)

Paper Project Page Code

NeurIPS

HEIR: Learning Graph-Based Motion Hierarchies

Cheng Zheng†, William Koch†, Baiang Li, Felix Heide († Equal contribution)

Paper Project Page Code

ICCP (Oral)

Sagiri: Low Dynamic Range Image Enhancement with Generative Diffusion Prior

Baiang Li, Sizhuo Ma, Yanhong Zeng, Xiaogang Xu, Youqing Fang, Zhao Zhang, Jian Wang*, Kai Chen* (*Corresponding authors)

Paper Project Page Code

SIGGRAPH

Matting by Generation

Zhixiang Wang, Baiang Li, Jian Wang*, Yu-Lun Liu, Jinwei Gu, Yung-Yu Chuang, Shin'ichi Satoh

Paper Code Project Page Video

TMM

Diving Deep into Regions: Exploiting Regional Information Transformer for Single Image Deraining

Baiang Li, Zhao Zhang*, Huan Zheng, Xiaogang Xu, Yanyan Wei, Jingyi Zhang, Jicong Fan, Meng Wang

Paper Code

ACM-MM

Dynamic Grouped Interaction Network for Low-Light Stereo Image Enhancement

Baiang Li, Huan Zheng, Zhao Zhang*, Yang Zhao, Zhongqiu Zhao, Haijun Zhang

Paper Code Project Page

Challenge Reports

Talks

Efficient Computer Vision in (Next-Generation) Cameras — Princeton University, NJ, USA
(Conference Oral) Exposure-Limited Image Enhancement with Generative Diffusion Prior — ICCP 2025, Toronto, Canada
Lucky HDR Merging — NextCam Group, Adobe Inc., CA, USA