Papers › Multi-level Wavelet-CNN for Image Restoration

Multi-level Wavelet-CNN for Image Restoration

18 May 2018arXiv:1805.07071archive 2025-07-28

Pengju Liu, Hongzhi Zhang, Kai Zhang, Liang Lin, WangMeng Zuo

The tradeoff between receptive field size and efficiency is a crucial issue in low level vision. Plain convolutional networks (CNNs) generally enlarge the receptive field at the expense of computational cost. Recently, dilated filtering has been adopted to address this issue. But it suffers from gridding effect, and the resulting receptive field is only a sparse sampling of input image with checkerboard patterns. In this paper, we present a novel multi-level wavelet CNN (MWCNN) model for better tradeoff between receptive field size and computational efficiency. With the modified U-Net architecture, wavelet transform is introduced to reduce the size of feature maps in the contracting subnetwork. Furthermore, another convolutional layer is further used to decrease the channels of feature maps. In the expanding subnetwork, inverse wavelet transform is then deployed to reconstruct the high resolution feature maps. Our MWCNN can also be explained as the generalization of dilated filtering and subsampling, and can be applied to many image restoration tasks. The experimental results clearly show the effectiveness of MWCNN for image denoising, single image super-resolution, and JPEG image artifacts removal.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

lpj0/MWCNN officialmentioned in paper report
MatusPilnan/nsiete-project mentioned on GitHubtf report
shaonianruntu/MWCNN mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Computational EfficiencyDenoisingImage DenoisingImage RestorationImage Super-ResolutionJPEG Artifact CorrectionSuper-Resolution

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Grayscale Image Denoising BSD68 sigma15 MWCNN PSNR 31.86 #6 of 16 Archive leaderboard report
Grayscale Image Denoising BSD68 sigma25 MWCNN PSNR 29.41 #4 of 16 Archive leaderboard report
Grayscale Image Denoising BSD68 sigma50 MWCNN PSNR 26.53 #4 of 15 Archive leaderboard report
Grayscale Image Denoising Set12 sigma15 MWCNN PSNR 33.15 #4 of 8 Archive leaderboard report
Grayscale Image Denoising Set12 sigma25 MWCNN PSNR 30.79 #3 of 6 Archive leaderboard report
Grayscale Image Denoising Set12 sigma50 MWCNN PSNR 27.74 #3 of 8 Archive leaderboard report
Grayscale Image Denoising Urban100 sigma15 MWCNN PSNR 33.17 #5 of 7 Archive leaderboard report
Grayscale Image Denoising Urban100 sigma25 MWCNN PSNR 30.66 #8 of 10 Archive leaderboard report
Grayscale Image Denoising Urban100 sigma50 MWCNN PSNR 27.42 #8 of 10 Archive leaderboard report
Image Super-Resolution BSD100 - 2x upscaling MWCNN PSNR 32.23 #21 of 30 Archive leaderboard report
Image Super-Resolution BSD100 - 3x upscaling MWCNN PSNR 29.12 #15 of 21 Archive leaderboard report
Image Super-Resolution BSD100 - 4x upscaling MWCNN PSNR 27.62 #30 of 71 Archive leaderboard report
Image Super-Resolution BSD100 - 4x upscaling MWCNN SSIM 0.7355 #30 of 71 Archive leaderboard report
Image Super-Resolution Set14 - 2x upscaling MWCNN PSNR 33.7 #22 of 35 Archive leaderboard report
Image Super-Resolution Set14 - 3x upscaling MWCNN PSNR 30.16 #18 of 24 Archive leaderboard report
Image Super-Resolution Set14 - 4x upscaling MWCNN PSNR 28.41 #67 of 104 Archive leaderboard report
Image Super-Resolution Set14 - 4x upscaling MWCNN SSIM 0.7816 #67 of 104 Archive leaderboard report
Image Super-Resolution Set5 - 2x upscaling MWCNN PSNR 37.91 #28 of 41 Archive leaderboard report
Image Super-Resolution Set5 - 3x upscaling MWCNN PSNR 34.17 #24 of 32 Archive leaderboard report
Image Super-Resolution Urban100 - 2x upscaling MWCNN PSNR 32.3 #23 of 29 Archive leaderboard report
Image Super-Resolution Urban100 - 3x upscaling MWCNN PSNR 28.13 #20 of 22 Archive leaderboard report
Image Super-Resolution Urban100 - 4x upscaling MWCNN PSNR 26.27 #40 of 65 Archive leaderboard report
Image Super-Resolution Urban100 - 4x upscaling MWCNN SSIM 0.7890 #40 of 65 Archive leaderboard report
JPEG Artifact Correction Classic5 (Quality 10 Grayscale) MWCNN PSNR 30.01 #4 of 7 Archive leaderboard report
JPEG Artifact Correction Classic5 (Quality 20 Grayscale) MWCNN PSNR 32.16 #4 of 6 Archive leaderboard report
JPEG Artifact Correction Classic5 (Quality 30 Grayscale) MWCNN PSNR 33.43 #4 of 6 Archive leaderboard report
JPEG Artifact Correction Classic5 (Quality 40 Grayscale) MWCNN PSNR 34.27 #4 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 10 Color) MWCNN PSNR 30.76 #5 of 6 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 10 Color) MWCNN PSNR-B 31.21 #5 of 6 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 10 Color) MWCNN SSIM 0.779 #5 of 6 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 10 Grayscale) MWCNN PSNR 34.12 #3 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 10 Grayscale) MWCNN PSNR-B 34.06 #3 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 10 Grayscale) MWCNN SSIM 0.884 #3 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 20 Color) MWCNN PSNR 32.79 #4 of 6 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 20 Color) MWCNN PSNR-B 33.32 #4 of 6 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 20 Color) MWCNN SSIM 0.812 #4 of 6 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 20 Grayscale) MWCNN PSNR 36.56 #2 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 20 Grayscale) MWCNN PSNR-B 36.44 #2 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 20 Grayscale) MWCNN SSIM 0.902 #2 of 5 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 30 Color) MWCNN PSNR 34.11 #3 of 4 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 30 Color) MWCNN PSNR-B 34.69 #3 of 4 Archive leaderboard report
JPEG Artifact Correction ICB (Quality 30 Color) MWCNN SSIM 0.845 #3 of 4 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 10 Color) MWCNN PSNR 27.45 #5 of 9 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 10 Color) MWCNN PSNR-B 27.44 #5 of 9 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 10 Color) MWCNN SSIM 0.808 #5 of 9 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 20 Color) MWCNN PSNR 29.80 #6 of 9 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 20 Color) MWCNN PSNR-B 29.78 #6 of 9 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 20 Color) MWCNN SSIM 0.877 #6 of 9 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 20 Grayscale) MWCNN PSNR 32.04 #5 of 12 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 20 Grayscale) MWCNN PSNR-B 31.83 #5 of 12 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 20 Grayscale) MWCNN SSIM 0.8989 #5 of 12 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 30 Grayscale) MWCNN PSNR 33.45 #4 of 7 Archive leaderboard report
JPEG Artifact Correction LIVE1 (Quality 40 Grayscale) MWCNN PSNR 34.45 #4 of 5 Archive leaderboard report
JPEG Artifact Correction Live1 (Quality 10 Grayscale) MWCNN PSNR 29.69 #5 of 13 Archive leaderboard report
JPEG Artifact Correction Live1 (Quality 10 Grayscale) MWCNN PSNR-B 29.39 #5 of 13 Archive leaderboard report
JPEG Artifact Correction Live1 (Quality 10 Grayscale) MWCNN SSIM 0.8357 #5 of 13 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Concatenated Skip ConnectionConvolutionMax PoolingReLUU-Net

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections