This is the PyTorch implementation of paper: Invertible Denoising Network: A Light Solution for Real Noise Removal (CVPR 2021). arxiv.
- Python 3 (Recommend to use Anaconda)
- PyTorch >= 1.0
- NVIDIA GPU + CUDA
- Python packages:
pip install numpy opencv-python lmdb pyyaml - TensorBoard:
- PyTorch >= 1.1:
pip install tb-nightly future - PyTorch == 1.0:
pip install tensorboardX
- PyTorch >= 1.1:
The datasets used in this paper is DND (can be downloaded here), SIDD (can be downloaded here) and RNI.
Training and testing codes are in 'codes/'. Please see 'codes/README.md' for basic usages. Pretrained model can be found in 'pretrained/'
All visual results for SIDD dataset can be found in 'results/'.
The code is based on Invertible Image Rescaling. If you find this code is helpful, please also cite the paper Invertible Image Rescaling.
If you find this work helps you, please cite:
@article{liu2021invertible,
title={Invertible Denoising Network: A Light Solution for Real Noise Removal},
author={Liu, Yang and Qin, Zhenyue and Anwar, Saeed and Ji, Pan and Kim, Dongwoo and Caldwell, Sabrina and Gedeon, Tom},
journal={arXiv preprint arXiv:2104.10546},
year={2021}
}
If you have any questions, please contact yang.liu3@anu.edu.au or zhenyue.qin@anu.edu.au.