{"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/checkerboard-artifact-free-sub-pixel","title":"Checkerboard artifact free sub-pixel convolution: A note on sub-pixel convolution, resize convolution and convolution resize","arxiv_id":"1707.02937","date":"2017-07-10","proceeding":null,"authors":["Andrew Aitken","Christian Ledig","Lucas Theis","Jose Caballero","Zehan Wang","Wenzhe Shi"],"abstract":"The most prominent problem associated with the deconvolution layer is the\npresence of checkerboard artifacts in output images and dense labels. To combat\nthis problem, smoothness constraints, post processing and different\narchitecture designs have been proposed. Odena et al. highlight three sources\nof checkerboard artifacts: deconvolution overlap, random initialization and\nloss functions. In this note, we proposed an initialization method for\nsub-pixel convolution known as convolution NN resize. Compared to sub-pixel\nconvolution initialized with schemes designed for standard convolution kernels,\nit is free from checkerboard artifacts immediately after initialization.\nCompared to resize convolution, at the same computational complexity, it has\nmore modelling power and converges to solutions with smaller test errors.","url_abs":"http://arxiv.org/abs/1707.02937v1","url_pdf":"http://arxiv.org/pdf/1707.02937v1.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":"checkerboard-artifact-free-sub-pixel","repo_url":"https://github.com/imatge-upc/3D-GAN-superresolution","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"checkerboard-artifact-free-sub-pixel","repo_url":"https://github.com/kostyaev/ICNR","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"checkerboard-artifact-free-sub-pixel","repo_url":"https://github.com/r06922019/butt_lion_paper_notes","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}