{"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/sequential-gating-ensemble-network-for-noise","title":"Sequential Gating Ensemble Network for Noise Robust Multi-Scale Face Restoration","arxiv_id":"1812.11834","date":"2018-12-19","proceeding":null,"authors":["Zhibo Chen","Jianxin Lin","Tiankuang Zhou","Feng Wu"],"abstract":"Face restoration from low resolution and noise is important for applications\nof face analysis recognition. However, most existing face restoration models\nomit the multiple scale issues in face restoration problem, which is still not\nwell-solved in research area. In this paper, we propose a Sequential Gating\nEnsemble Network (SGEN) for multi-scale noise robust face restoration issue. To\nendow the network with multi-scale representation ability, we first employ the\nprinciple of ensemble learning for SGEN network architecture designing. The\nSGEN aggregates multi-level base-encoders and base-decoders into the network,\nwhich enables the network to contain multiple scales of receptive field.\nInstead of combining these base-en/decoders directly with non-sequential\noperations, the SGEN takes base-en/decoders from different levels as sequential\ndata. Specifically, it is visualized that SGEN learns to sequentially extract\nhigh level information from base-encoders in bottom-up manner and restore low\nlevel information from base-decoders in top-down manner. Besides, we propose to\nrealize bottom-up and top-down information combination and selection with\nSequential Gating Unit (SGU). The SGU sequentially takes information from two\ndifferent levels as inputs and decides the output based on one active input.\nExperiment results on benchmark dataset demonstrate that our SGEN is more\neffective at multi-scale human face restoration with more image details and\nless noise than state-of-the-art image restoration models. Further utilizing\nadversarial training scheme, SGEN also produces more visually preferred results\nthan other models under subjective evaluation.","url_abs":"http://arxiv.org/abs/1812.11834v1","url_pdf":"http://arxiv.org/pdf/1812.11834v1.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":"sequential-gating-ensemble-network-for-noise","repo_url":"https://github.com/tomorrowi6/SGEN-pytorch","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"ensemble-learning","task_name":"Ensemble Learning"},{"task_slug":"image-restoration","task_name":"Image Restoration"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}