NIRN: Self-supervised Noisy Image Reconstruction Network for Real-World Image Denoising
- Ubuntu 16.04
- CUDA 10.0 & cuDNN 7.5.0
- Python 3.7.6
Environment configuration:
conda env create -f NIRN.yml
conda activate NIRN
The package comprises these functions：
*) single_free_denoise.py : Denoise single-frame real-world noisy image
*) multi_free_denoise.py : Denoise multi-frame real-world noisy images（frame = 4）
*) single_real_image_denoise.py : Denoise single-frame real-world noisy image， we provide noise-free image to get PSNR result
*) multi_real_image_denoise.py : Denoise multi-frame real-world noisy images（frame = 4）， we provide noise-free image to get PSNR result
*) single_Gauss_denoise.py : Denoise single-frame image with Gaussian noise (sigma = 50)
*) multi_Gauss_denoise.py : Denoise multi-frame image with Gaussian noise (sigma = 50) （frame = 4）
Implementation Details: Here we save the output from both subnetworks IGM and NGM with the suffixes '_CIGM_denoised.png' and '_NGM_denoised.png' When changing the picture, it is better to use the name of our picture, otherwise an error may occur.