{"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-fashion-compatibility-with","title":"Learning Fashion Compatibility with Bidirectional LSTMs","arxiv_id":"1707.05691","date":"2017-07-18","proceeding":null,"authors":["Xintong Han","Zuxuan Wu","Yu-Gang Jiang","Larry S. Davis"],"abstract":"The ubiquity of online fashion shopping demands effective recommendation\nservices for customers. In this paper, we study two types of fashion\nrecommendation: (i) suggesting an item that matches existing components in a\nset to form a stylish outfit (a collection of fashion items), and (ii)\ngenerating an outfit with multimodal (images/text) specifications from a user.\nTo this end, we propose to jointly learn a visual-semantic embedding and the\ncompatibility relationships among fashion items in an end-to-end fashion. More\nspecifically, we consider a fashion outfit to be a sequence (usually from top\nto bottom and then accessories) and each item in the outfit as a time step.\nGiven the fashion items in an outfit, we train a bidirectional LSTM (Bi-LSTM)\nmodel to sequentially predict the next item conditioned on previous ones to\nlearn their compatibility relationships. Further, we learn a visual-semantic\nspace by regressing image features to their semantic representations aiming to\ninject attribute and category information as a regularization for training the\nLSTM. The trained network can not only perform the aforementioned\nrecommendations effectively but also predict the compatibility of a given\noutfit. We conduct extensive experiments on our newly collected Polyvore\ndataset, and the results provide strong qualitative and quantitative evidence\nthat our framework outperforms alternative methods.","url_abs":"http://arxiv.org/abs/1707.05691v1","url_pdf":"http://arxiv.org/pdf/1707.05691v1.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-fashion-compatibility-with","repo_url":"https://github.com/AemikaChow/DATASOURCE","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"learning-fashion-compatibility-with","repo_url":"https://github.com/xthan/polyvore","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[{"slug":"polyvore","name":"Polyvore","full_name":"Polyvore Outfits"}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.05691","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}