【ICCV'2023🔥】Implicit Neural Representation for Cooperative Low-light Image Enhancement

Welcome! This is the official implementation of our paper: Implicit Neural Representation for Cooperative Low-light Image Enhancement

Authors: Shuzhou Yang, Moxuan Ding, Yanmin Wu, Zihan Li, Jian Zhang*.

📣 News

(2023.7.17): Our code has been released❗️
(2023.7.14): 🎉🎉🎉 Our paper has been accepted to ICCV 2023❗️

Overview

Prerequisites

Linux or macOS
Python 3.8
NVIDIA GPU + CUDA CuDNN

🔑 Setup

Type the command:

pip install -r requirements.txt

🧩 Download

You need create a directory ./saves/[YOUR-MODEL] (e.g., ./saves/LSRW).
Download the pre-trained models and put them into ./saves/[YOUR-MODEL].
Here we release two versions of the pre-trained model, which are trained on LSRW and LOL datasets respectively:

🚀 Quick Run

Create directories ./dataset/testA and ./dataset/testB. Put your test images in ./dataset/testA (And you should keep whatever one image in ./dataset/testB to make sure program can start.)
Test the model with the pre-trained weights:

CUDA_VISIBLE_DEVICES=0 python test.py --dataroot ./dataset --name [YOUR-MODEL] --preprocess=none

The test results will be saved to a directory here: ./results/[YOUR-MODEL]/test_latest/images, and will also be displayed in an html file here: ./results/[YOUR-MODEL]/test_latest/index.html.

🤖 Training

Download training low-light data and put it in ./dataset/trainA.
Randomly adopt hundreds of normal-light images and put them in ./dataset/trainB.
Train a model:

cd NeRCo-main
mkdir loss
CUDA_VISIBLE_DEVICES=0 python train.py --dataroot ./dataset --name [YOUR-MODEL]

Loss curve can be found in the directory ./loss.
To see more intermediate results, check out ./saves/[YOUR-MODEL]/web/index.html.

📌 Citation

If you find this code useful for your research, please use the following BibTeX entry.

@InProceedings{Yang_2023_ICCV,
    author    = {Yang, Shuzhou and Ding, Moxuan and Wu, Yanmin and Li, Zihan and Zhang, Jian},
    title     = {Implicit Neural Representation for Cooperative Low-light Image Enhancement},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2023},
    pages     = {12918-12927}
}

Ysz2022/NeRCo