ELVIS: Enhance Low-Light for Video Instance Segmentation in the Dark [arXiv preprint]

Authors: Joanne Lin, Ruirui Lin, Yini Li, David Bull, Nantheera Anantrasirichai

Institution: Visual Information Laboratory, University of Bristol, United Kingdom

[Project Page][arXiv]

Usage

1. Train VDP-Net
2. Collect $\Phi$ from low-light datasets
3. Train Mask2Former-based VIS method with ELVIS (we provide GenVIS+ELVIS code as an example)

We also provide the pre-trained weights of VDP-Net and GenVIS+ELVIS here.

Citation

If you use our work in your research, please cite using the following BibTeX entry:

@article{lin2025elvis,
         title={ELVIS: Enhance Low-Light for Video Instance Segmentation in the Dark},
         author={Lin, Joanne and Lin, Ruirui, and Li, Yini and Bull, David and Anatrasirichai, Nantheera},
         year={2025},
         publisher={arXiv}}