{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/multi-level-wavelet-cnn-for-image-restoration","title":"Multi-level Wavelet-CNN for Image Restoration","arxiv_id":"1805.07071","date":"2018-05-18","proceeding":null,"authors":["Pengju Liu","Hongzhi Zhang","Kai Zhang","Liang Lin","WangMeng Zuo"],"abstract":"The tradeoff between receptive field size and efficiency is a crucial issue\nin low level vision. Plain convolutional networks (CNNs) generally enlarge the\nreceptive field at the expense of computational cost. Recently, dilated\nfiltering has been adopted to address this issue. But it suffers from gridding\neffect, and the resulting receptive field is only a sparse sampling of input\nimage with checkerboard patterns. In this paper, we present a novel multi-level\nwavelet CNN (MWCNN) model for better tradeoff between receptive field size and\ncomputational efficiency. With the modified U-Net architecture, wavelet\ntransform is introduced to reduce the size of feature maps in the contracting\nsubnetwork. Furthermore, another convolutional layer is further used to\ndecrease the channels of feature maps. In the expanding subnetwork, inverse\nwavelet transform is then deployed to reconstruct the high resolution feature\nmaps. Our MWCNN can also be explained as the generalization of dilated\nfiltering and subsampling, and can be applied to many image restoration tasks.\nThe experimental results clearly show the effectiveness of MWCNN for image\ndenoising, single image super-resolution, and JPEG image artifacts removal.","url_abs":"http://arxiv.org/abs/1805.07071v2","url_pdf":"http://arxiv.org/pdf/1805.07071v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"multi-level-wavelet-cnn-for-image-restoration","repo_url":"https://github.com/lpj0/MWCNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"multi-level-wavelet-cnn-for-image-restoration","repo_url":"https://github.com/MatusPilnan/nsiete-project","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"multi-level-wavelet-cnn-for-image-restoration","repo_url":"https://github.com/Shakib-IO/Diminishing_Image_Noise_Using_Deep_Learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"multi-level-wavelet-cnn-for-image-restoration","repo_url":"https://github.com/chintan1995/Image-Denoising-using-Deep-Learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"multi-level-wavelet-cnn-for-image-restoration","repo_url":"https://github.com/shaonianruntu/MWCNN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"},{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"image-denoising","task_name":"Image Denoising"},{"task_slug":"image-restoration","task_name":"Image Restoration"},{"task_slug":"image-super-resolution","task_name":"Image Super-Resolution"},{"task_slug":"jpeg-artifact-correction","task_name":"JPEG Artifact Correction"},{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"u-net","method_name":"U-Net"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/grayscale-image-denoising-on-bsd68-sigma15","task":"Grayscale Image Denoising","dataset":"BSD68 sigma15","model":"MWCNN","rank_in_archive_order":6,"of":16,"metrics":{"PSNR":"31.86"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-bsd68-sigma25","task":"Grayscale Image Denoising","dataset":"BSD68 sigma25","model":"MWCNN","rank_in_archive_order":4,"of":16,"metrics":{"PSNR":"29.41"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-bsd68-sigma50","task":"Grayscale Image Denoising","dataset":"BSD68 sigma50","model":"MWCNN","rank_in_archive_order":4,"of":15,"metrics":{"PSNR":"26.53"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-set12-sigma15","task":"Grayscale Image Denoising","dataset":"Set12 sigma15","model":"MWCNN","rank_in_archive_order":4,"of":8,"metrics":{"PSNR":"33.15"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-set12-sigma25","task":"Grayscale Image Denoising","dataset":"Set12 sigma25","model":"MWCNN","rank_in_archive_order":3,"of":6,"metrics":{"PSNR":"30.79"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-set12-sigma50","task":"Grayscale Image Denoising","dataset":"Set12 sigma50","model":"MWCNN","rank_in_archive_order":3,"of":8,"metrics":{"PSNR":"27.74"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-urban100-sigma15","task":"Grayscale Image Denoising","dataset":"Urban100 sigma15","model":"MWCNN","rank_in_archive_order":5,"of":7,"metrics":{"PSNR":"33.17"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-urban100-sigma25","task":"Grayscale Image Denoising","dataset":"Urban100 sigma25","model":"MWCNN","rank_in_archive_order":8,"of":10,"metrics":{"PSNR":"30.66"},"uses_additional_data":false},{"leaderboard":"/sota/grayscale-image-denoising-on-urban100-sigma50","task":"Grayscale Image Denoising","dataset":"Urban100 sigma50","model":"MWCNN","rank_in_archive_order":8,"of":10,"metrics":{"PSNR":"27.42"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-bsd100-2x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 2x upscaling","model":"MWCNN","rank_in_archive_order":21,"of":30,"metrics":{"PSNR":"32.23"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-bsd100-3x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 3x upscaling","model":"MWCNN","rank_in_archive_order":15,"of":21,"metrics":{"PSNR":"29.12"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-bsd100-4x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 4x upscaling","model":"MWCNN","rank_in_archive_order":30,"of":71,"metrics":{"PSNR":"27.62","SSIM":"0.7355"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-2x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 2x upscaling","model":"MWCNN","rank_in_archive_order":22,"of":35,"metrics":{"PSNR":"33.7"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-3x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 3x upscaling","model":"MWCNN","rank_in_archive_order":18,"of":24,"metrics":{"PSNR":"30.16"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"MWCNN","rank_in_archive_order":67,"of":104,"metrics":{"PSNR":"28.41","SSIM":"0.7816"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-2x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 2x upscaling","model":"MWCNN","rank_in_archive_order":28,"of":41,"metrics":{"PSNR":"37.91"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-3x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 3x upscaling","model":"MWCNN","rank_in_archive_order":24,"of":32,"metrics":{"PSNR":"34.17"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-2x","task":"Image Super-Resolution","dataset":"Urban100 - 2x upscaling","model":"MWCNN","rank_in_archive_order":23,"of":29,"metrics":{"PSNR":"32.3"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-3x","task":"Image Super-Resolution","dataset":"Urban100 - 3x upscaling","model":"MWCNN","rank_in_archive_order":20,"of":22,"metrics":{"PSNR":"28.13"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"MWCNN","rank_in_archive_order":40,"of":65,"metrics":{"PSNR":"26.27","SSIM":"0.7890"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-classic5-quality","task":"JPEG Artifact Correction","dataset":"Classic5 (Quality 10 Grayscale)","model":"MWCNN","rank_in_archive_order":4,"of":7,"metrics":{"PSNR":"30.01"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-classic5-quality-1","task":"JPEG Artifact Correction","dataset":"Classic5 (Quality 20 Grayscale)","model":"MWCNN","rank_in_archive_order":4,"of":6,"metrics":{"PSNR":"32.16"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-classic5-quality-2","task":"JPEG Artifact Correction","dataset":"Classic5 (Quality 30 Grayscale)","model":"MWCNN","rank_in_archive_order":4,"of":6,"metrics":{"PSNR":"33.43"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-classic5-quality-3","task":"JPEG Artifact Correction","dataset":"Classic5 (Quality 40 Grayscale)","model":"MWCNN","rank_in_archive_order":4,"of":5,"metrics":{"PSNR":"34.27"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-icb-quality-10","task":"JPEG Artifact Correction","dataset":"ICB (Quality 10 Color)","model":"MWCNN","rank_in_archive_order":5,"of":6,"metrics":{"PSNR":"30.76","PSNR-B":"31.21","SSIM":"0.779"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-icb-quality-10-1","task":"JPEG Artifact Correction","dataset":"ICB (Quality 10 Grayscale)","model":"MWCNN","rank_in_archive_order":3,"of":5,"metrics":{"PSNR":"34.12","PSNR-B":"34.06","SSIM":"0.884"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-icb-quality-20","task":"JPEG Artifact Correction","dataset":"ICB (Quality 20 Color)","model":"MWCNN","rank_in_archive_order":4,"of":6,"metrics":{"PSNR":"32.79","PSNR-B":"33.32","SSIM":"0.812"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-icb-quality-20-1","task":"JPEG Artifact Correction","dataset":"ICB (Quality 20 Grayscale)","model":"MWCNN","rank_in_archive_order":2,"of":5,"metrics":{"PSNR":"36.56","PSNR-B":"36.44","SSIM":"0.902"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-icb-quality-30","task":"JPEG Artifact Correction","dataset":"ICB (Quality 30 Color)","model":"MWCNN","rank_in_archive_order":3,"of":4,"metrics":{"PSNR":"34.11","PSNR-B":"34.69","SSIM":"0.845"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-live1-quality-10","task":"JPEG Artifact Correction","dataset":"LIVE1 (Quality 10 Color)","model":"MWCNN","rank_in_archive_order":5,"of":9,"metrics":{"PSNR":"27.45","PSNR-B":"27.44","SSIM":"0.808"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-live1-quality-20","task":"JPEG Artifact Correction","dataset":"LIVE1 (Quality 20 Color)","model":"MWCNN","rank_in_archive_order":6,"of":9,"metrics":{"PSNR":"29.80","PSNR-B":"29.78","SSIM":"0.877"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-live1-quality-20-1","task":"JPEG Artifact Correction","dataset":"LIVE1 (Quality 20 Grayscale)","model":"MWCNN","rank_in_archive_order":5,"of":12,"metrics":{"PSNR":"32.04","PSNR-B":"31.83","SSIM":"0.8989"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-live1-quality-30-1","task":"JPEG Artifact Correction","dataset":"LIVE1 (Quality 30 Grayscale)","model":"MWCNN","rank_in_archive_order":4,"of":7,"metrics":{"PSNR":"33.45"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-live1-quality-40","task":"JPEG Artifact Correction","dataset":"LIVE1 (Quality 40 Grayscale)","model":"MWCNN","rank_in_archive_order":4,"of":5,"metrics":{"PSNR":"34.45"},"uses_additional_data":false},{"leaderboard":"/sota/jpeg-artifact-correction-on-live1-quality-10-1","task":"JPEG Artifact Correction","dataset":"Live1 (Quality 10 Grayscale)","model":"MWCNN","rank_in_archive_order":5,"of":13,"metrics":{"PSNR":"29.69","PSNR-B":"29.39","SSIM":"0.8357"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1805.07071","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}