{"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/iterative-residual-cnns-for-burst-photography","title":"Iterative Residual CNNs for Burst Photography Applications","arxiv_id":"1811.12197","date":"2018-11-29","proceeding":"CVPR 2019 6","authors":["Filippos Kokkinos","Stamatios Lefkimmiatis"],"abstract":"Modern inexpensive imaging sensors suffer from inherent hardware constraints\nwhich often result in captured images of poor quality. Among the most common\nways to deal with such limitations is to rely on burst photography, which\nnowadays acts as the backbone of all modern smartphone imaging applications. In\nthis work, we focus on the fact that every frame of a burst sequence can be\naccurately described by a forward (physical) model. This in turn allows us to\nrestore a single image of higher quality from a sequence of low quality images\nas the solution of an optimization problem. Inspired by an extension of the\ngradient descent method that can handle non-smooth functions, namely the\nproximal gradient descent, and modern deep learning techniques, we propose a\nconvolutional iterative network with a transparent architecture. Our network,\nuses a burst of low quality image frames and is able to produce an output of\nhigher image quality recovering fine details which are not distinguishable in\nany of the original burst frames. We focus both on the burst photography\npipeline as a whole, i.e. burst demosaicking and denoising, as well as on the\ntraditional Gaussian denoising task. The developed method demonstrates\nconsistent state-of-the art performance across the two tasks and as opposed to\nother recent deep learning approaches does not have any inherent restrictions\neither to the number of frames or their ordering. Code can be found at\nhttps://fkokkinos.github.io/deep_burst/","url_abs":"http://arxiv.org/abs/1811.12197v2","url_pdf":"http://arxiv.org/pdf/1811.12197v2.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":"iterative-residual-cnns-for-burst-photography","repo_url":"https://github.com/cig-skoltech/burst-cvpr-2019","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"demosaicking","task_name":"Demosaicking"},{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"image-restoration","task_name":"Image Restoration"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.12197","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}