{"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/residual-non-local-attention-networks-for-1","title":"Residual Non-local Attention Networks for Image Restoration","arxiv_id":"1903.10082","date":"2019-03-24","proceeding":"ICLR 2019 5","authors":["Yulun Zhang","Kunpeng Li","Kai Li","Bineng Zhong","Yun Fu"],"abstract":"In this paper, we propose a residual non-local attention network for\nhigh-quality image restoration. Without considering the uneven distribution of\ninformation in the corrupted images, previous methods are restricted by local\nconvolutional operation and equal treatment of spatial- and channel-wise\nfeatures. To address this issue, we design local and non-local attention blocks\nto extract features that capture the long-range dependencies between pixels and\npay more attention to the challenging parts. Specifically, we design trunk\nbranch and (non-)local mask branch in each (non-)local attention block. The\ntrunk branch is used to extract hierarchical features. Local and non-local mask\nbranches aim to adaptively rescale these hierarchical features with mixed\nattentions. The local mask branch concentrates on more local structures with\nconvolutional operations, while non-local attention considers more about\nlong-range dependencies in the whole feature map. Furthermore, we propose\nresidual local and non-local attention learning to train the very deep network,\nwhich further enhance the representation ability of the network. Our proposed\nmethod can be generalized for various image restoration applications, such as\nimage denoising, demosaicing, compression artifacts reduction, and\nsuper-resolution. Experiments demonstrate that our method obtains comparable or\nbetter results compared with recently leading methods quantitatively and\nvisually.","url_abs":"http://arxiv.org/abs/1903.10082v1","url_pdf":"http://arxiv.org/pdf/1903.10082v1.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":"residual-non-local-attention-networks-for-1","repo_url":"https://github.com/yulunzhang/RNAN","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"residual-non-local-attention-networks-for-1","repo_url":"https://github.com/bruinxiong/RNAN","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-denoising","task_name":"Image Denoising"},{"task_slug":"image-restoration","task_name":"Image Restoration"},{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1903.10082","atlas_url":"https://app.syntology.ai/?focus=1903.10082","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}