This repository contains the official code to reproduce the results from the paper:
CVF-SID: Cyclic multi-Variate Function for Self-Supervised Image Denoising by Disentangling Noise from Image (CVPR 2022)
[arXiv] [presentation]
Clone this repository into any place you want.
git clone https://github.com/Reyhanehne/CVF-SID_PyTorch.git
cd CVF-SID_PyTorch
To train and evaluate the model directly please visit SIDD website or Drive and download the original Noisy sRGB data
and Ground-truth sRGB data
from SIDD Validation Data and Ground Truth
and place them in data/SIDD_Small_sRGB_Only
folder.
Download config.json
and model_best.pth
from this link and save them in models/CVF_SID/SIDD_Val/
folder.
You can now go to src folder and test our CVF-SID by:
python test.py --device 0 --config ../models/CVF_SID/SIDD_Val/config.json --resume ../models/CVF_SID/SIDD_Val/model_best.pth
or you can train it by yourself as follows:
python train.py --device 0 --config config_SIDD_Val.json --tag SIDD_Val
If you find our code or paper useful, please consider citing:
@inproceedings{Neshatavar2022CVFSIDCM,
title={CVF-SID: Cyclic multi-Variate Function for Self-Supervised Image Denoising by Disentangling Noise from Image},
author={Reyhaneh Neshatavar and Mohsen Yavartanoo and Sanghyun Son and Kyoung Mu Lee},
booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2022}
}