library_name: pytorch
license: other
pipeline_tag: image-to-video
tags:
- brdfusion
- inverse-rendering
- relighting
- urban-scenes
- waymo-open-dataset
BRDFusion Pretrained Checkpoints
This repository contains the pretrained checkpoints for the paper BRDFusion: Physics Meets Generation for Urban Scene Inverse Rendering.
Project Page | GitHub | arXiv
BRDFusion is a unified framework that combines physics-based inverse rendering with generative modeling for high-quality urban scene relighting. It decomposes urban videos into geometry, materials, and HDR lighting, enabling applications such as novel-view rendering, night simulation, and dynamic object insertion/editing.
By combining these two complementary approaches, BRDFusion recovers explicit, consistent scene properties with physical modeling and alleviates optimization ambiguity with generative priors. During forward rendering, the physical model provides controllable rendering from the scene configuration, and the generative model denoises and fixes artifacts.
Citation
@misc{liu2026brdfusionphysicsmeetsgeneration,
title={BRDFusion: Physics Meets Generation for Urban Scene Inverse Rendering},
author={Yi-Ruei Liu and Jie-Ying Lee and Zheng-Hui Huang and Yu-Lun Liu and Chih-Hao Lin},
year={2026},
eprint={2606.17049},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2606.17049},
}