{"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/hifacegan-face-renovation-via-collaborative","title":"HiFaceGAN: Face Renovation via Collaborative Suppression and Replenishment","arxiv_id":"2005.05005","date":"2020-05-11","proceeding":null,"authors":["Lingbo Yang","Chang Liu","Pan Wang","Shanshe Wang","Peiran Ren","Siwei Ma","Wen Gao"],"abstract":"Existing face restoration researches typically relies on either the degradation prior or explicit guidance labels for training, which often results in limited generalization ability over real-world images with heterogeneous degradations and rich background contents. In this paper, we investigate the more challenging and practical \"dual-blind\" version of the problem by lifting the requirements on both types of prior, termed as \"Face Renovation\"(FR). Specifically, we formulated FR as a semantic-guided generation problem and tackle it with a collaborative suppression and replenishment (CSR) approach. This leads to HiFaceGAN, a multi-stage framework containing several nested CSR units that progressively replenish facial details based on the hierarchical semantic guidance extracted from the front-end content-adaptive suppression modules. Extensive experiments on both synthetic and real face images have verified the superior performance of HiFaceGAN over a wide range of challenging restoration subtasks, demonstrating its versatility, robustness and generalization ability towards real-world face processing applications.","url_abs":"https://arxiv.org/abs/2005.05005v2","url_pdf":"https://arxiv.org/pdf/2005.05005v2.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":"hifacegan-face-renovation-via-collaborative","repo_url":"https://github.com/Lotayou/Face-Renovation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"GPL-3.0"}},{"paper_slug":"hifacegan-face-renovation-via-collaborative","repo_url":"https://github.com/2023-MindSpore-1/ms-code-214/tree/main/HiFaceGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"hifacegan-face-renovation-via-collaborative","repo_url":"https://github.com/2023-MindSpore-4/Code-5/tree/main/HiFaceGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"hifacegan-face-renovation-via-collaborative","repo_url":"https://github.com/Mind23-2/MindCode-3/tree/main/HiFaceGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"hifacegan-face-renovation-via-collaborative","repo_url":"https://github.com/MindSpore-paper-code-3/code4/tree/main/HiFaceGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"blind-face-restoration","task_name":"Blind Face Restoration"},{"task_slug":"face-hallucination","task_name":"Face Hallucination"},{"task_slug":"image-denoising","task_name":"Image Denoising"},{"task_slug":"image-restoration","task_name":"Image Restoration"},{"task_slug":"image-super-resolution","task_name":"Image Super-Resolution"}],"methods":[{"method_slug":"adaptive-instance-normalization","method_name":"Adaptive Instance Normalization"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/blind-face-restoration-on-celeba-test","task":"Blind Face Restoration","dataset":"CelebA-Test","model":"HiFaceGAN","rank_in_archive_order":10,"of":15,"metrics":{"Deg.":"42.18","FID":"66.09","LPIPS":"47.7","NIQE":"4.916","PSNR":"24.92","SSIM":"0.6195"},"uses_additional_data":false},{"leaderboard":"/sota/face-hallucination-on-ffhq-512-x-512-16x","task":"Face Hallucination","dataset":"FFHQ 512 x 512 - 16x upscaling","model":"HiFaceGAN","rank_in_archive_order":1,"of":4,"metrics":{"FID":"11.389","LPIPS":"0.2449","NIQE":"6.767"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-ffhq-1024-x-1024-4x","task":"Image Super-Resolution","dataset":"FFHQ 1024 x 1024 - 4x upscaling","model":"HiFaceGAN","rank_in_archive_order":1,"of":9,"metrics":{"FID":"1.978","MS-SSIM":"0.975","PSNR":"33.04","SSIM":"0.875"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-ffhq-256-x-256-4x","task":"Image Super-Resolution","dataset":"FFHQ 256 x 256 - 4x upscaling","model":"HiFaceGAN","rank_in_archive_order":1,"of":11,"metrics":{"FID":"5.36","MS-SSIM":"0.971","PSNR":"28.65","SSIM":"0.816"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-ffhq-512-x-512-4x","task":"Image Super-Resolution","dataset":"FFHQ 512 x 512 - 4x upscaling","model":"HiFaceGAN","rank_in_archive_order":1,"of":8,"metrics":{"FED":"0.0716","FID":"1.898","LLE":"2.071","LPIPS":"0.0723","MS-SSIM":"0.971","NIQE":"6.961","PSNR":"30.824","SSIM":"0.838"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2005.05005","atlas_url":"https://app.syntology.ai/?focus=2005.05005","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}