{"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/deep-image-demosaicking-using-a-cascade-of","title":"Deep Image Demosaicking using a Cascade of Convolutional Residual Denoising Networks","arxiv_id":"1803.05215","date":"2018-03-14","proceeding":"ECCV 2018 9","authors":["Filippos Kokkinos","Stamatios Lefkimmiatis"],"abstract":"Demosaicking and denoising are among the most crucial steps of modern digital\ncamera pipelines and their joint treatment is a highly ill-posed inverse\nproblem where at-least two-thirds of the information are missing and the rest\nare corrupted by noise. This poses a great challenge in obtaining meaningful\nreconstructions and a special care for the efficient treatment of the problem\nis required. While there are several machine learning approaches that have been\nrecently introduced to deal with joint image demosaicking-denoising, in this\nwork we propose a novel deep learning architecture which is inspired by\npowerful classical image regularization methods and large-scale convex\noptimization techniques. Consequently, our derived network is more transparent\nand has a clear interpretation compared to alternative competitive deep\nlearning approaches. Our extensive experiments demonstrate that our network\noutperforms any previous approaches on both noisy and noise-free data. This\nimprovement in reconstruction quality is attributed to the principled way we\ndesign our network architecture, which also requires fewer trainable parameters\nthan the current state-of-the-art deep network solution. Finally, we show that\nour network has the ability to generalize well even when it is trained on small\ndatasets, while keeping the overall number of trainable parameters low.","url_abs":"http://arxiv.org/abs/1803.05215v4","url_pdf":"http://arxiv.org/pdf/1803.05215v4.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":"deep-image-demosaicking-using-a-cascade-of","repo_url":"https://github.com/cig-skoltech/deep_demosaick","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"demosaicking","task_name":"Demosaicking"},{"task_slug":"denoising","task_name":"Denoising"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1803.05215","atlas_url":"https://app.syntology.ai/?focus=1803.05215","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}