AGLLNet: Attention Guided Low-light Image Enhancement (IJCV 2021)

This is the test code for “Attention Guided Low-light Image Enhancement with a Large Scale Low-light Simulation Dataset” in IJCV 2021, by Feifan Lv, Yu Li, and Feng Lu.

Paper | ArXiv | Project page (data)

Requirements

• python 3.5

• Tensorflow 1.6.0

• Keras 2.2.0

• imageio

• opencv

Usage

Testing

You can put you image into the folder input and run

cd AGLLNet
python run_agllnet.py

The results will be stored in the folder output.

Training:

Training code will NOT be provided this time.

Model

• AgLLNet.h5 (This model is newly trained for general low light enhancement. It is not strictly the one used in our IJCV paper).

Bibtex

If you use this code for your research, please consider star this repo and cite our paper.

@article{lv2021attention,
 title={Attention guided low-light image enhancement with a large scale low-light simulation dataset},
 author={Lv, Feifan and Li, Yu and Lu, Feng},
 journal={International Journal of Computer Vision},
 volume={129},
 number={7},
 pages={2175--2193},
 year={2021}
}

Related work: stable low light video enhancement

Learning Temporal Consistency for Low Light Video Enhancement from Single Images (CVPR2021) Paper | Code