FFDNet_pytorch

• A PyTorch implementation of a denoising network called FFDNet
• Paper: FFDNet: Toward a Fast and Flexible Solution for CNN based Image Denoising - arxiv / IEEE

Dataset

• Waterloo Exploration Database

Usage

• Train

python3 ffdnet.py \
    --use_gpu \
    --is_train \
    --train_path './train_data/' \
    --model_path './models/' \
    --batch_size 768 \
    --epoches 80 \
    --val_epoch 5
    --patch_size 32 \
    --save_checkpoints 20 \
    --train_noise_interval 15 75 15 \
    --val_noise_interval 30 60 30 \

• Test

python3 ffdnet.py \
    --use_gpu \
    --is_test \
    --test_path './test_data/color.png' \
    --model_path './models/' \
    --add_noise
    --noise_sigma 30

References

• Some codes are copied from An Analysis and Implementation of the FFDNet Image Denoising Method