Event-based Low-illumination Image Enhancement (TMM)
Welcome to the repository for our project on Event-based Low-illumination Image Enhancement.
We provide the following pre-trained models for your use:
- Indoor model: indoor_model
- Outdoor model: outdoor_model
- Total model: total_model
The LIE Dataset used in our research can be downloaded from data.
To test the performance of our models, you can use the provided code.
Once you have configured the specific paths, you can run the following code. The pre-trained models should be placed in the pretrained directory, while the datasets and scenarios need to be stored in the data directory.
The visualized results and metrics will be saved in the result directory.
# Example code for testing
python Test_Ours.py --resume ./pretrained/model_indoor.pth --data ./data/LIEDataset/orig_indoor_test --save ./result/display_indoor/If you wish to train on your own data, please read and execute the train.py script, and modify the corresponding parameter settings.
More detailed parameter configurations can be found in the config file.
The codebase for this project is built upon the foundation of pytorch-template and Restormer's work.
If you find our work useful, please consider citing our paper:
@ARTICLE{10168206,
author={Jiang, Yu and Wang, Yuehang and Li, Siqi and Zhang, Yongji and Zhao, Minghao and Gao, Yue},
journal={IEEE Transactions on Multimedia},
title={Event-Based Low-Illumination Image Enhancement},
year={2024},
volume={26},
number={},
pages={1920-1931},
doi={10.1109/TMM.2023.3290432}}