Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising

13 Aug 2016Kai ZhangWangmeng ZuoYunjin ChenDeyu MengLei Zhang

Discriminative model learning for image denoising has been recently attracting considerable attentions due to its favorable denoising performance. In this paper, we take one step forward by investigating the construction of feed-forward denoising convolutional neural networks (DnCNNs) to embrace the progress in very deep architecture, learning algorithm, and regularization method into image denoising... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Image Super-Resolution BSD100 - 2x upscaling DnCNN-3 PSNR 31.9 # 11
Image Super-Resolution BSD100 - 3x upscaling DnCNN-3 PSNR 28.85 # 8
Image Super-Resolution BSD100 - 4x upscaling DnCNN-3 PSNR 27.29 # 25
SSIM 0.7253 # 30
Color Image Denoising BSD68 sigma15 DnCNN-3 PSNR 31.46 # 4
Grayscale Image Denoising BSD68 sigma25 DnCNN PSNR 29.23 # 5
Color Image Denoising BSD68 sigma25 DnCNN-3 PSNR 29.02 # 4
Color Image Denoising BSD68 sigma50 DnCNN-3 PSNR 26.1 # 5
Color Image Denoising CBSD68 sigma35 DnCNN-B* PSNR 28.74 # 4
Image Compression Artifact Reduction Classic5 Quality 10 DnCNN-3 PSNR 29.4 # 4
Image Compression Artifact Reduction Classic5 Quality 20 DnCNN-3 PSNR 31.63 # 3
Image Compression Artifact Reduction Classic5 Quality 30 DnCNN-3 PSNR 32.91 # 3
Image Compression Artifact Reduction Classic5 Quality 40 DnCNN-3 PSNR 33.77 # 3
Image Compression Artifact Reduction LIVE1 Quality 10 DnCNN-3 PSNR 29.19 # 5
Image Compression Artifact Reduction LIVE1 Quality 20 DnCNN-3 PSNR 31.59 # 5
Image Compression Artifact Reduction LIVE1 Quality 30 DnCNN-3 PSNR 32.98 # 3
Image Compression Artifact Reduction LIVE1 Quality 40 DnCNN-3 PSNR 33.96 # 3
Image Super-Resolution Set14 - 2x upscaling DnCNN-3 PSNR 33.03 # 12
Image Super-Resolution Set14 - 3x upscaling DnCNN-3 PSNR 29.81 # 6
Image Super-Resolution Set14 - 4x upscaling DnCNN-3 PSNR 28.04 # 30
SSIM 0.7672 # 34
Image Super-Resolution Set5 - 2x upscaling DnCNN-3 PSNR 37.58 # 12
Image Super-Resolution Set5 - 3x upscaling DnCNN-3 PSNR 33.75 # 10
Image Super-Resolution Set5 - 4x upscaling DnCNN-3 PSNR 31.4 # 27
SSIM 0.8845 # 28
Image Super-Resolution Urban100 - 2x upscaling DnCNN-3 PSNR 30.74 # 11
Image Super-Resolution Urban100 - 3x upscaling DnCNN-3 PSNR 27.15 # 6
Image Super-Resolution Urban100 - 4x upscaling DnCNN-3 PSNR 25.2 # 30
SSIM 0.7521 # 26
Grayscale Image Denoising Urban100 sigma15 DnCNN PSNR 32.67 # 4
Grayscale Image Denoising Urban100 sigma25 DnCNN PSNR 29.97 # 5

Results from Other Papers


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK SOURCE PAPER COMPARE
Denoising Darmstadt Noise Dataset CDnCNN-B PSNR 32.43 # 7