{"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/learning-linear-transformations-for-fast","title":"Learning Linear Transformations for Fast Arbitrary Style Transfer","arxiv_id":"1808.04537","date":"2018-08-14","proceeding":null,"authors":["Xueting Li","Sifei Liu","Jan Kautz","Ming-Hsuan Yang"],"abstract":"Given a random pair of images, an arbitrary style transfer method extracts\nthe feel from the reference image to synthesize an output based on the look of\nthe other content image. Recent arbitrary style transfer methods transfer\nsecond order statistics from reference image onto content image via a\nmultiplication between content image features and a transformation matrix,\nwhich is computed from features with a pre-determined algorithm. These\nalgorithms either require computationally expensive operations, or fail to\nmodel the feature covariance and produce artifacts in synthesized images.\nGeneralized from these methods, in this work, we derive the form of\ntransformation matrix theoretically and present an arbitrary style transfer\napproach that learns the transformation matrix with a feed-forward network. Our\nalgorithm is highly efficient yet allows a flexible combination of multi-level\nstyles while preserving content affinity during style transfer process. We\ndemonstrate the effectiveness of our approach on four tasks: artistic style\ntransfer, video and photo-realistic style transfer as well as domain\nadaptation, including comparisons with the state-of-the-art methods.","url_abs":"http://arxiv.org/abs/1808.04537v1","url_pdf":"http://arxiv.org/pdf/1808.04537v1.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":"learning-linear-transformations-for-fast","repo_url":"https://github.com/sunshineatnoon/LinearStyleTransfer","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"style-transfer","task_name":"Style Transfer"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.04537","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}