DAM-VSR: Disentanglement of Appearance and Motion for Video Super-Resolution

Zhe Kong1,2,3, Le Li4, Yong Zhang2, Feng Gao2, Shaoshu Yang2,5, Tao Wang6, Kaihao Zhang7, Zhuoliang Kang2, Xiaoming Wei2, Guanying Chen1, Wenhan Luo3
1Shenzhen Campus of Sun Yat-sen University, 2Meituan, 3The Hong Kong University of Science and Technology, 4Tianjin University, 5School of Artificial Intelligence, University of Chinese Academy of Sciences, 6Nanjing University, 7Harbin Institute of Technology, Shenzhen
Corresponding Author


ACM SIGGRAPH 2025

Paper arXiv Video Code

AIGC Videos (upscale ×4)

Real-World Videos (upscale ×4)

Abstract

Real-world video super-resolution (VSR) presents significant challenges due to complex and unpredictable degradations. Although some recent methods utilize image diffusion models for VSR and have shown improved detail generation capabilities, they still struggle to produce temporally consistent frames. We attempt to use Stable Video Diffusion (SVD) combined with ControlNet to address this issue. However, due to the intrinsic image-animation characteristics of SVD, it is challenging to generate fine details using only low-quality videos. To tackle this problem, we propose DAM-VSR, an appearance and motion disentanglement framework for VSR. This framework disentangles VSR into appearance enhancement and motion control problems. Specifically, appearance enhancement is achieved through reference image super-resolution, while motion control is achieved through video ControlNet. This disentanglement fully leverages the generative prior of video diffusion models and the detail generation capabilities of image super-resolution models. Furthermore, equipped with the proposed motion-aligned bidirectional sampling strategy, DAM-VSR can conduct VSR on longer input videos. DAM-VSR achieves state-of-the-art performance on real-world data and AIGC data, demonstrating its powerful detail generation capabilities.

Method

In this work, we propose DAM-VSR, an appearance and motion disentanglement framework for VSR. To support long video generation, we propose a motion-aligned bidirectional sampling strategy, which consists of a disentangled forward generation process and a disentangled backward generation process. These two processes maintain temporal consistency through motion alignment.

BibTeX

@inproceedings{kong2025dam,
      title={DAM-VSR: Disentanglement of Appearance and Motion for Video Super-Resolution}, 
      author={Zhe Kong, Le Li, Yong Zhang, Feng Gao, Shaoshu Yang, Tao Wang, Kaihao Zhang, Zhuoliang Kang, Xiaoming Wei, Guanying Chen, Wenhan Luo},
      year={2025},
      booktitle={ACM SIGGRAPH 2025},
}