{"url":"/sota/image-denoising-on-sidd","task":{"name":"Image Denoising","url":"/task/image-denoising","note":null},"dataset":{"name":"SIDD","url":"/dataset/sidd"},"category":"Computer Vision","categories":["Computer Vision","Medical"],"category_note":null,"description":"**Image Denoising** is a computer vision task that involves removing noise from an image. Noise can be introduced into an image during acquisition or processing, and can reduce image quality and make it difficult to interpret. Image denoising techniques aim to restore an image to its original quality by reducing or removing the noise, while preserving the important features of the image.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [Wide Inference Network for Image Denoising via\r\nLearning Pixel-distribution Prior](https://arxiv.org/pdf/1707.05414v5.pdf) )</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["PSNR (sRGB)","SSIM (sRGB)","Average PSNR"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"PSNR (sRGB)":"higher","SSIM (sRGB)":"higher","Average PSNR":"higher"}},"counts":{"rows":22,"rows_with_code":22,"rows_with_paper_page":22,"rows_dated":22,"rows_using_additional_data":7},"rows":[{"rank_in_archive_order":1,"model":"CGNet","metrics":{"PSNR (sRGB)":"40.39","SSIM (sRGB)":"0.964"},"uses_additional_data":false,"paper_date":"2024-01-26","paper":"/paper/cascadedgaze-efficiency-in-global-context","paper_url":"https://arxiv.org/abs/2401.15235v2","paper_title":"CascadedGaze: Efficiency in Global Context Extraction for Image Restoration","code":"https://github.com/Ascend-Research/CascadedGaze","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":1,"n_samples":11,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"KBNet","metrics":{"PSNR (sRGB)":"40.35","SSIM (sRGB)":"0.972"},"uses_additional_data":false,"paper_date":"2023-03-06","paper":"/paper/kbnet-kernel-basis-network-for-image","paper_url":"https://arxiv.org/abs/2303.02881v1","paper_title":"KBNet: Kernel Basis Network for Image Restoration","code":"https://github.com/zhangyi-3/kbnet","n_code_links":1,"syntology":{"n_ran":6,"n_unverified":3,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"NAFNet","metrics":{"PSNR (sRGB)":"40.30","SSIM (sRGB)":"0.961"},"uses_additional_data":false,"paper_date":"2022-04-10","paper":"/paper/simple-baselines-for-image-restoration","paper_url":"https://arxiv.org/abs/2204.04676v4","paper_title":"Simple Baselines for Image Restoration","code":"https://github.com/megvii-research/NAFNet","n_code_links":13,"syntology":{"n_ran":23,"n_unverified":7,"n_samples":30,"n_pointer_only_licence":17}},{"rank_in_archive_order":4,"model":"SSAMAN","metrics":{"PSNR (sRGB)":"40.08","SSIM (sRGB)":"0.962"},"uses_additional_data":false,"paper_date":"2024-04-10","paper":"/paper/single-stage-adaptive-multi-attention-network","paper_url":"https://ieeexplore.ieee.org/abstract/document/10495777","paper_title":"Single Stage Adaptive Multi-Attention Network for Image Restoration","code":"https://github.com/anas-zafar/SSAMAN","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"PNGAN","metrics":{"PSNR (sRGB)":"40.07","SSIM (sRGB)":"0.960"},"uses_additional_data":true,"paper_date":"2022-04-06","paper":"/paper/learning-to-generate-realistic-noisy-images-2","paper_url":"https://arxiv.org/abs/2204.02844v2","paper_title":"Learning to Generate Realistic Noisy Images via Pixel-level Noise-aware Adversarial Training","code":"https://github.com/caiyuanhao1998/PNGAN","n_code_links":2,"syntology":{"n_ran":14,"n_unverified":4,"n_samples":18,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"Restormer","metrics":{"PSNR (sRGB)":"40.02","SSIM (sRGB)":"0.960"},"uses_additional_data":false,"paper_date":"2021-11-18","paper":"/paper/restormer-efficient-transformer-for-high","paper_url":"https://arxiv.org/abs/2111.09881v2","paper_title":"Restormer: Efficient Transformer for High-Resolution Image Restoration","code":"https://github.com/swz30/restormer","n_code_links":13,"syntology":{"n_ran":4,"n_unverified":0,"n_samples":4,"n_pointer_only_licence":4}},{"rank_in_archive_order":7,"model":"HINet","metrics":{"PSNR (sRGB)":"39.99","SSIM (sRGB)":"0.958"},"uses_additional_data":false,"paper_date":"2021-05-13","paper":"/paper/hinet-half-instance-normalization-network-for","paper_url":"https://arxiv.org/abs/2105.06086v2","paper_title":"HINet: Half Instance Normalization Network for Image Restoration","code":"https://github.com/megvii-model/HINet","n_code_links":2,"syntology":null},{"rank_in_archive_order":8,"model":"MAXIM-3S","metrics":{"PSNR (sRGB)":"39.96","SSIM (sRGB)":"0.960"},"uses_additional_data":false,"paper_date":"2022-01-09","paper":"/paper/maxim-multi-axis-mlp-for-image-processing","paper_url":"https://arxiv.org/abs/2201.02973v2","paper_title":"MAXIM: Multi-Axis MLP for Image Processing","code":"https://github.com/google-research/maxim","n_code_links":3,"syntology":{"n_ran":31,"n_unverified":15,"n_samples":46,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"Uformer-B","metrics":{"PSNR (sRGB)":"39.89","SSIM (sRGB)":"0.960"},"uses_additional_data":false,"paper_date":"2021-06-06","paper":"/paper/uformer-a-general-u-shaped-transformer-for","paper_url":"https://arxiv.org/abs/2106.03106v2","paper_title":"Uformer: A General U-Shaped Transformer for Image Restoration","code":"https://github.com/ZhendongWang6/Uformer","n_code_links":4,"syntology":{"n_ran":6,"n_unverified":0,"n_samples":6,"n_pointer_only_licence":2}},{"rank_in_archive_order":10,"model":"MIRNet","metrics":{"PSNR (sRGB)":"39.72","SSIM (sRGB)":"0.959"},"uses_additional_data":false,"paper_date":"2020-03-15","paper":"/paper/learning-enriched-features-for-real-image","paper_url":"https://arxiv.org/abs/2003.06792v2","paper_title":"Learning Enriched Features for Real Image Restoration and Enhancement","code":"https://github.com/swz30/restormer","n_code_links":12,"syntology":{"n_ran":12,"n_unverified":8,"n_samples":20,"n_pointer_only_licence":3}},{"rank_in_archive_order":11,"model":"SRMNet","metrics":{"PSNR (sRGB)":"39.72","SSIM (sRGB)":"0.959"},"uses_additional_data":false,"paper_date":"2022-03-03","paper":"/paper/selective-residual-m-net-for-real-image","paper_url":"https://arxiv.org/abs/2203.01645v1","paper_title":"Selective Residual M-Net for Real Image Denoising","code":"https://github.com/FanChiMao/SRMNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"MPRNet","metrics":{"PSNR (sRGB)":"39.71","SSIM (sRGB)":"0.958"},"uses_additional_data":false,"paper_date":"2021-02-04","paper":"/paper/multi-stage-progressive-image-restoration","paper_url":"https://arxiv.org/abs/2102.02808v2","paper_title":"Multi-Stage Progressive Image Restoration","code":"https://github.com/swz30/restormer","n_code_links":8,"syntology":{"n_ran":18,"n_unverified":8,"n_samples":26,"n_pointer_only_licence":25}},{"rank_in_archive_order":13,"model":"AKDT","metrics":{"PSNR (sRGB)":"39.70","SSIM (sRGB)":"0.961"},"uses_additional_data":false,"paper_date":"2025-02-26","paper":"/paper/akdt-adaptive-kernel-dilation-transformer-for","paper_url":"https://www.insticc.org/node/TechnicalProgram/VISIGRAPP/2025/presentationDetails/131577","paper_title":"AKDT: Adaptive Kernel Dilation Transformer for Effective Image Denoising","code":"https://github.com/albrateanu/AKDT","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"CycleISP","metrics":{"PSNR (sRGB)":"39.52","SSIM (sRGB)":"0.957"},"uses_additional_data":true,"paper_date":"2020-03-17","paper":"/paper/cycleisp-real-image-restoration-via-improved","paper_url":"https://arxiv.org/abs/2003.07761v1","paper_title":"CycleISP: Real Image Restoration via Improved Data Synthesis","code":"https://github.com/swz30/restormer","n_code_links":8,"syntology":null},{"rank_in_archive_order":15,"model":"DANet+","metrics":{"PSNR (sRGB)":"39.47","SSIM (sRGB)":"0.957"},"uses_additional_data":true,"paper_date":"2020-07-12","paper":"/paper/dual-adversarial-network-toward-real-world","paper_url":"https://arxiv.org/abs/2007.05946v1","paper_title":"Dual Adversarial Network: Toward Real-world Noise Removal and Noise Generation","code":"https://github.com/zsyOAOA/DANet","n_code_links":2,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":1}},{"rank_in_archive_order":16,"model":"SADNet","metrics":{"PSNR (sRGB)":"39.46","SSIM (sRGB)":"0.957"},"uses_additional_data":true,"paper_date":"2020-01-28","paper":"/paper/spatial-adaptive-network-for-single-image","paper_url":"https://arxiv.org/abs/2001.10291v2","paper_title":"Spatial-Adaptive Network for Single Image Denoising","code":"https://github.com/JimmyChame/SADNet","n_code_links":2,"syntology":null},{"rank_in_archive_order":17,"model":"VDN","metrics":{"PSNR (sRGB)":"39.28","SSIM (sRGB)":"0.956"},"uses_additional_data":false,"paper_date":"2019-08-29","paper":"/paper/variational-denoising-network-toward-blind","paper_url":"https://arxiv.org/abs/1908.11314v5","paper_title":"Variational Denoising Network: Toward Blind Noise Modeling and Removal","code":"https://github.com/zsyOAOA/VDNet","n_code_links":2,"syntology":null},{"rank_in_archive_order":18,"model":"AINDNet","metrics":{"PSNR (sRGB)":"38.95","SSIM (sRGB)":"0.952"},"uses_additional_data":true,"paper_date":"2020-02-26","paper":"/paper/transfer-learning-from-synthetic-to-real-2","paper_url":"https://arxiv.org/abs/2002.11244v2","paper_title":"Transfer Learning from Synthetic to Real-Noise Denoising with Adaptive Instance Normalization","code":"https://github.com/terryoo/AINDNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":19,"model":"RIDNet","metrics":{"PSNR (sRGB)":"38.71","SSIM (sRGB)":"0.951"},"uses_additional_data":true,"paper_date":"2019-04-16","paper":"/paper/real-image-denoising-with-feature-attention","paper_url":"https://arxiv.org/abs/1904.07396v2","paper_title":"Real Image Denoising with Feature Attention","code":"https://github.com/saeed-anwar/RIDNet","n_code_links":3,"syntology":null},{"rank_in_archive_order":20,"model":"CBDNet","metrics":{"PSNR (sRGB)":"30.78","SSIM (sRGB)":"0.801"},"uses_additional_data":true,"paper_date":"2018-07-12","paper":"/paper/toward-convolutional-blind-denoising-of-real","paper_url":"http://arxiv.org/abs/1807.04686v2","paper_title":"Toward Convolutional Blind Denoising of Real Photographs","code":"https://github.com/GuoShi28/CBDNet","n_code_links":3,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":21,"model":"NBNet","metrics":{"SSIM (sRGB)":"0.973"},"uses_additional_data":false,"paper_date":"2020-12-30","paper":"/paper/nbnet-noise-basis-learning-for-image","paper_url":"https://arxiv.org/abs/2012.15028v2","paper_title":"NBNet: Noise Basis Learning for Image Denoising with Subspace Projection","code":"https://github.com/megengine/nbnet","n_code_links":9,"syntology":null},{"rank_in_archive_order":22,"model":"DRANet","metrics":{"Average PSNR":"39.50","SSIM (sRGB)":"0.957"},"uses_additional_data":false,"paper_date":"2023-05-07","paper":"/paper/dual-residual-attention-network-for-image","paper_url":"https://arxiv.org/abs/2305.04269v1","paper_title":"Dual Residual Attention Network for Image Denoising","code":"https://github.com/WenCongWu/DRANet","n_code_links":1,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":11,"rows_with_any_sample_ran":10,"distinct_papers_with_graph_line":11,"distinct_papers_with_any_sample_ran":10,"samples_over_distinct_papers":{"n_ran":125,"n_unverified":47,"n_samples":172,"n_pointer_only_licence":52,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":125,"n_unverified":47,"n_samples":172,"n_pointer_only_licence":52,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}