{"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/context-patch-face-hallucination-based-on","title":"Context-Patch Face Hallucination Based on Thresholding Locality-constrained Representation and Reproducing Learning","arxiv_id":"1809.00665","date":"2018-09-03","proceeding":null,"authors":["Junjun Jiang","Yi Yu","Suhua Tang","Jiayi Ma","Akiko Aizawa","Kiyoharu Aizawa"],"abstract":"Face hallucination is a technique that reconstruct high-resolution (HR) faces\nfrom low-resolution (LR) faces, by using the prior knowledge learned from HR/LR\nface pairs. Most state-of-the-arts leverage position-patch prior knowledge of\nhuman face to estimate the optimal representation coefficients for each image\npatch. However, they focus only the position information and usually ignore the\ncontext information of image patch. In addition, when they are confronted with\nmisalignment or the Small Sample Size (SSS) problem, the hallucination\nperformance is very poor. To this end, this study incorporates the contextual\ninformation of image patch and proposes a powerful and efficient context-patch\nbased face hallucination approach, namely Thresholding Locality-constrained\nRepresentation and Reproducing learning (TLcR-RL). Under the context-patch\nbased framework, we advance a thresholding based representation method to\nenhance the reconstruction accuracy and reduce the computational complexity. To\nfurther improve the performance of the proposed algorithm, we propose a\npromotion strategy called reproducing learning. By adding the estimated HR face\nto the training set, which can simulates the case that the HR version of the\ninput LR face is present in the training set, thus iteratively enhancing the\nfinal hallucination result. Experiments demonstrate that the proposed TLcR-RL\nmethod achieves a substantial increase in the hallucinated results, both\nsubjectively and objectively. Additionally, the proposed framework is more\nrobust to face misalignment and the SSS problem, and its hallucinated HR face\nis still very good when the LR test face is from the real-world. The MATLAB\nsource code is available at https://github.com/junjun-jiang/TLcR-RL","url_abs":"http://arxiv.org/abs/1809.00665v2","url_pdf":"http://arxiv.org/pdf/1809.00665v2.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":"context-patch-face-hallucination-based-on","repo_url":"https://github.com/junjun-jiang/Face-Hallucination-Benchmark","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"context-patch-face-hallucination-based-on","repo_url":"https://github.com/junjun-jiang/TLcR-RL","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"face-hallucination","task_name":"Face Hallucination"},{"task_slug":"hallucination","task_name":"Hallucination"},{"task_slug":null,"task_name":"Position"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}