nveu's Introduction

NIR-assisted Low-light Video Enhancement Using Unpaired 24-hour Data

In this repository, we provide source codes and pretrained checkpoint for run the test demo.

FMSVD Dataset

You can download the FMSVD dataset here (split) or here (one zip file).

Third-Party Dataset and NIR Relighting Codes

You can dowload the third-party dataset and NIR relighting codes from this link.

Environment Setup

conda create -n nir python==3.8
conda activate nir
conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c pytorch -c conda-forge
python -m pip install detectron2 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html
pip install opencv-python
cd package_core
python setup.py install
cd ..
pip install timm

Download pretrained checkpoint

Download checkpoints from here, and put it into ./.

Run Test Script

chmod 777 test.sh
./test.sh

The results will be saved to ./test_results.

nveu's People

Contributors

Stargazers

Watchers

nveu's Issues

Inquiry About Code for Generating Pseudo-Nighttime NIR and VIS Images from Daytime NIR and VIS Images

Hello，
I could not find the specific code or methodology for generating pseudo-nighttime NIR and VIS images from daytime NIR and VIS images. This aspect of your research is particularly intriguing to me, and I believe it could significantly enhance my current work.

Could you kindly provide guidance or share the relevant code for this process? Any additional details or references you could provide would be immensely appreciated.

Thank you very much for your time and assistance. I look forward to your response.

Recommend Projects