{"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/fsrnet-end-to-end-learning-face-super","title":"FSRNet: End-to-End Learning Face Super-Resolution with Facial Priors","arxiv_id":"1711.10703","date":"2017-11-29","proceeding":"CVPR 2018 6","authors":["Yu Chen","Ying Tai","Xiaoming Liu","Chunhua Shen","Jian Yang"],"abstract":"Face Super-Resolution (SR) is a domain-specific super-resolution problem. The\nspecific facial prior knowledge could be leveraged for better super-resolving\nface images. We present a novel deep end-to-end trainable Face Super-Resolution\nNetwork (FSRNet), which makes full use of the geometry prior, i.e., facial\nlandmark heatmaps and parsing maps, to super-resolve very low-resolution (LR)\nface images without well-aligned requirement. Specifically, we first construct\na coarse SR network to recover a coarse high-resolution (HR) image. Then, the\ncoarse HR image is sent to two branches: a fine SR encoder and a prior\ninformation estimation network, which extracts the image features, and\nestimates landmark heatmaps/parsing maps respectively. Both image features and\nprior information are sent to a fine SR decoder to recover the HR image. To\nfurther generate realistic faces, we propose the Face Super-Resolution\nGenerative Adversarial Network (FSRGAN) to incorporate the adversarial loss\ninto FSRNet. Moreover, we introduce two related tasks, face alignment and\nparsing, as the new evaluation metrics for face SR, which address the\ninconsistency of classic metrics w.r.t. visual perception. Extensive benchmark\nexperiments show that FSRNet and FSRGAN significantly outperforms state of the\narts for very LR face SR, both quantitatively and qualitatively. Code will be\nmade available upon publication.","url_abs":"http://arxiv.org/abs/1711.10703v1","url_pdf":"http://arxiv.org/pdf/1711.10703v1.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":"fsrnet-end-to-end-learning-face-super","repo_url":"https://github.com/tyshiwo/FSRNet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"fsrnet-end-to-end-learning-face-super","repo_url":"https://github.com/Dou-Yu-xuan/FSRNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"fsrnet-end-to-end-learning-face-super","repo_url":"https://github.com/ZoieMo/Multi-task","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"fsrnet-end-to-end-learning-face-super","repo_url":"https://github.com/cs-giung/FSRNet-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"face-alignment","task_name":"Face Alignment"},{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1711.10703","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}