Wakeup-Darkness

[ACM TOMM] Wakeup-Darkness: When Multimodal Meets Unsupervised Low-light Image Enhancement

The structure

The grad-cam results about CSFF

Compare results of LOL dataset

Model

weights/ model.py fuse_block.py

Dataset processing

you should run the sam.py to produce the segment image like LOL/LIME/SCIE

Train and Test

python train.py --arc WithoutCalNet --batch_size 10

python test.py

python test_clip.py

Citation

@article{zhang2025,
  title={Wakeup-Darkness: When Multimodal Meets Unsupervised Low-light Image Enhancement},
  author={Xiaofeng Zhang, Zishan Xu, Hao Tang, Chaochen Gu, Wei Chen},
  journal={ACM Transactions on Multimedia Computing, Communications and Applications},
  year={2025}
}