{"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/convolutional-neural-network-with-median","title":"Convolutional Neural Network with Median Layers for Denoising Salt-and-Pepper Contaminations","arxiv_id":"1908.06452","date":"2019-08-18","proceeding":null,"authors":["Luming Liang","Sen Deng","Lionel Gueguen","Mingqiang Wei","Xinming Wu","Jing Qin"],"abstract":"We propose a deep fully convolutional neural network with a new type of layer, named median layer, to restore images contaminated by the salt-and-pepper (s&p) noise. A median layer simply performs median filtering on all feature channels. By adding this kind of layer into some widely used fully convolutional deep neural networks, we develop an end-to-end network that removes the extremely high-level s&p noise without performing any non-trivial preprocessing tasks, which is different from all the existing literature in s&p noise removal. Experiments show that inserting median layers into a simple fully-convolutional network with the L2 loss significantly boosts the signal-to-noise ratio. Quantitative comparisons testify that our network outperforms the state-of-the-art methods with a limited amount of training data. The source code has been released for public evaluation and use (https://github.com/llmpass/medianDenoise).","url_abs":"https://arxiv.org/abs/1908.06452v1","url_pdf":"https://arxiv.org/pdf/1908.06452v1.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":"convolutional-neural-network-with-median","repo_url":"https://github.com/llmpass/medianDenoise","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"salt-and-pepper-noise-removal","task_name":"Salt-And-Pepper Noise Removal"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/salt-and-pepper-noise-removal-on-bsd300-noise","task":"Salt-And-Pepper Noise Removal","dataset":"BSD300 Noise Level 30%","model":"CNN (Median Layers)","rank_in_archive_order":2,"of":3,"metrics":{"PSNR":"40.90"},"uses_additional_data":false},{"leaderboard":"/sota/salt-and-pepper-noise-removal-on-bsd300-noise-1","task":"Salt-And-Pepper Noise Removal","dataset":"BSD300 Noise Level 50%","model":"CNN (Median Layers)","rank_in_archive_order":2,"of":3,"metrics":{"PSNR":"37.28"},"uses_additional_data":false},{"leaderboard":"/sota/salt-and-pepper-noise-removal-on-bsd300-noise-2","task":"Salt-And-Pepper Noise Removal","dataset":"BSD300 Noise Level 70%","model":"CNN (Median Layers)","rank_in_archive_order":2,"of":3,"metrics":{"PSNR":"32.4"},"uses_additional_data":false},{"leaderboard":"/sota/salt-and-pepper-noise-removal-on-kodak24-1","task":"Salt-And-Pepper Noise Removal","dataset":"Kodak24 Noise Level 30%","model":"CNN (Median Layers)","rank_in_archive_order":1,"of":3,"metrics":{"PSNR":"36.39"},"uses_additional_data":false},{"leaderboard":"/sota/salt-and-pepper-noise-removal-on-kodak24-2","task":"Salt-And-Pepper Noise Removal","dataset":"Kodak24 Noise Level 50%","model":"CNN (Median Layers)","rank_in_archive_order":1,"of":3,"metrics":{"PSNR":"34.35"},"uses_additional_data":false},{"leaderboard":"/sota/salt-and-pepper-noise-removal-on-kodak24-3","task":"Salt-And-Pepper Noise Removal","dataset":"Kodak24 Noise Level 70%","model":"CNN (Median Layers)","rank_in_archive_order":1,"of":3,"metrics":{"PSNR":"31.56"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}