Deep Burst Denoising

Unofficial implement of Deep Burst Denoising
https://arxiv.org/abs/1712.05790

Enviroment

requirement.txt

Dataset

Use SIDD dataset.
Have two folder : noisy image and ground true image

Input folder have struct :

/
    /noise
        /[scene_instance]
            /[image].PNG
    /gt
        /[scene_instance]
            /[image].PNG

Train

Train Single image

python train.py -n /home/dell/Downloads/FullTest/noisy -g /home/dell/Downloads/FullTest/clean -sz 256 -nw 8 -bs 2 -ep 100 -se 100 --type single -r SFD_C_99.pth.tar

Train Multi image

python train.py -n /home/dell/Downloads/FullTest/noisy -g /home/dell/Downloads/FullTest/clean -sz 256 -nw 8 -bs 2 -ep 100 -se 100 --type multi -r MFD_C_99.pth.tar

References

[1] https://arxiv.org/abs/1712.05790 , Godard, Clément, Kevin Matzen, and Matt Uyttendaele. “Deep burst denoising.” Proceedings of the European Conference on Computer Vision (ECCV). 2018.

[2] https://github.com/Ourshanabi/Burst-denoising