{"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/a-unified-framework-for-generalizable-style","title":"A Unified Framework for Generalizable Style Transfer: Style and Content Separation","arxiv_id":"1806.05173","date":"2018-06-13","proceeding":null,"authors":["Yexun Zhang","Ya zhang","Wenbin Cai"],"abstract":"Image style transfer has drawn broad attention in recent years. However, most\nexisting methods aim to explicitly model the transformation between different\nstyles, and the learned model is thus not generalizable to new styles. We here\npropose a unified style transfer framework for both character typeface transfer\nand neural style transfer tasks leveraging style and content separation. A key\nmerit of such framework is its generalizability to new styles and contents. The\noverall framework consists of style encoder, content encoder, mixer and\ndecoder. The style encoder and content encoder are used to extract the style\nand content representations from the corresponding reference images. The mixer\nintegrates the above two representations and feeds it into the decoder to\ngenerate images with the target style and content. During training, the encoder\nnetworks learn to extract styles and contents from limited size of\nstyle/content reference images. This learning framework allows simultaneous\nstyle transfer among multiple styles and can be deemed as a special\n`multi-task' learning scenario. The encoders are expected to capture the\nunderlying features for different styles and contents which is generalizable to\nnew styles and contents. Under this framework, we design two individual\nnetworks for character typeface transfer and neural style transfer,\nrespectively. For character typeface transfer, to separate the style features\nand content features, we leverage the conditional dependence of styles and\ncontents given an image. For neural style transfer, we leverage the statistical\ninformation of feature maps in certain layers to represent style. Extensive\nexperimental results have demonstrated the effectiveness and robustness of the\nproposed methods.","url_abs":"http://arxiv.org/abs/1806.05173v1","url_pdf":"http://arxiv.org/pdf/1806.05173v1.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":"a-unified-framework-for-generalizable-style","repo_url":"https://github.com/Ign0reLee/EMD_Neural_Style_Transfer_with_Tensorflow","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"multi-task-learning","task_name":"Multi-Task Learning"},{"task_slug":"style-transfer","task_name":"Style Transfer"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.05173","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}