{"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/multigrid-backprojection-super-resolution-and","title":"Multigrid Backprojection Super-Resolution and Deep Filter Visualization","arxiv_id":"1809.09326","date":"2018-09-25","proceeding":null,"authors":["Pablo Navarrete Michelini","Hanwen Liu","Dan Zhu"],"abstract":"We introduce a novel deep-learning architecture for image upscaling by large\nfactors (e.g. 4x, 8x) based on examples of pristine high-resolution images. Our\ntarget is to reconstruct high-resolution images from their downscale versions.\nThe proposed system performs a multi-level progressive upscaling, starting from\nsmall factors (2x) and updating for higher factors (4x and 8x). The system is\nrecursive as it repeats the same procedure at each level. It is also residual\nsince we use the network to update the outputs of a classic upscaler. The\nnetwork residuals are improved by Iterative Back-Projections (IBP) computed in\nthe features of a convolutional network. To work in multiple levels we extend\nthe standard back-projection algorithm using a recursion analogous to\nMulti-Grid algorithms commonly used as solvers of large systems of linear\nequations. We finally show how the network can be interpreted as a standard\nupsampling-and-filter upscaler with a space-variant filter that adapts to the\ngeometry. This approach allows us to visualize how the network learns to\nupscale. Finally, our system reaches state of the art quality for models with\nrelatively few number of parameters.","url_abs":"http://arxiv.org/abs/1809.09326v3","url_pdf":"http://arxiv.org/pdf/1809.09326v3.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":"multigrid-backprojection-super-resolution-and","repo_url":"https://github.com/pnavarre/pirm-sr-2018","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}