{"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/gated-gan-adversarial-gated-networks-for","title":"Gated-GAN: Adversarial Gated Networks for Multi-Collection Style Transfer","arxiv_id":"1904.02296","date":"2019-04-04","proceeding":null,"authors":["Xinyuan Chen","Chang Xu","Xiaokang Yang","Li Song","DaCheng Tao"],"abstract":"Style transfer describes the rendering of an image semantic content as\ndifferent artistic styles. Recently, generative adversarial networks (GANs)\nhave emerged as an effective approach in style transfer by adversarially\ntraining the generator to synthesize convincing counterfeits. However,\ntraditional GAN suffers from the mode collapse issue, resulting in unstable\ntraining and making style transfer quality difficult to guarantee. In addition,\nthe GAN generator is only compatible with one style, so a series of GANs must\nbe trained to provide users with choices to transfer more than one kind of\nstyle. In this paper, we focus on tackling these challenges and limitations to\nimprove style transfer. We propose adversarial gated networks (Gated GAN) to\ntransfer multiple styles in a single model. The generative networks have three\nmodules: an encoder, a gated transformer, and a decoder. Different styles can\nbe achieved by passing input images through different branches of the gated\ntransformer. To stabilize training, the encoder and decoder are combined as an\nautoencoder to reconstruct the input images. The discriminative networks are\nused to distinguish whether the input image is a stylized or genuine image. An\nauxiliary classifier is used to recognize the style categories of transferred\nimages, thereby helping the generative networks generate images in multiple\nstyles. In addition, Gated GAN makes it possible to explore a new style by\ninvestigating styles learned from artists or genres. Our extensive experiments\ndemonstrate the stability and effectiveness of the proposed model for\nmultistyle transfer.","url_abs":"http://arxiv.org/abs/1904.02296v1","url_pdf":"http://arxiv.org/pdf/1904.02296v1.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":"gated-gan-adversarial-gated-networks-for","repo_url":"https://github.com/colemiller94/gatedgan","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"gated-gan-adversarial-gated-networks-for","repo_url":"https://github.com/xinyuanc91/Gated-GAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"style-transfer","task_name":"Style Transfer"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1904.02296","atlas_url":"https://app.syntology.ai/?focus=1904.02296","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}