{"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/aesthetic-driven-image-enhancement-by","title":"Aesthetic-Driven Image Enhancement by Adversarial Learning","arxiv_id":"1707.05251","date":"2017-07-17","proceeding":null,"authors":["Yubin Deng","Chen Change Loy","Xiaoou Tang"],"abstract":"We introduce EnhanceGAN, an adversarial learning based model that performs\nautomatic image enhancement. Traditional image enhancement frameworks typically\ninvolve training models in a fully-supervised manner, which require expensive\nannotations in the form of aligned image pairs. In contrast to these\napproaches, our proposed EnhanceGAN only requires weak supervision (binary\nlabels on image aesthetic quality) and is able to learn enhancement operators\nfor the task of aesthetic-based image enhancement. In particular, we show the\neffectiveness of a piecewise color enhancement module trained with weak\nsupervision, and extend the proposed EnhanceGAN framework to learning a deep\nfiltering-based aesthetic enhancer. The full differentiability of our image\nenhancement operators enables the training of EnhanceGAN in an end-to-end\nmanner. We further demonstrate the capability of EnhanceGAN in learning\naesthetic-based image cropping without any groundtruth cropping pairs. Our\nweakly-supervised EnhanceGAN reports competitive quantitative results on\naesthetic-based color enhancement as well as automatic image cropping, and a\nuser study confirms that our image enhancement results are on par with or even\npreferred over professional enhancement.","url_abs":"http://arxiv.org/abs/1707.05251v2","url_pdf":"http://arxiv.org/pdf/1707.05251v2.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":"aesthetic-driven-image-enhancement-by","repo_url":"https://github.com/dannysdeng/EnhanceGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"torch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"image-cropping","task_name":"Image Cropping"},{"task_slug":"image-enhancement","task_name":"Image Enhancement"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.05251","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}