{"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/sigan-siamese-generative-adversarial-network","title":"SiGAN: Siamese Generative Adversarial Network for Identity-Preserving Face Hallucination","arxiv_id":"1807.08370","date":"2018-07-22","proceeding":null,"authors":["Chih-Chung Hsu","Chia-Wen Lin","Weng-Tai Su","Gene Cheung"],"abstract":"Despite generative adversarial networks (GANs) can hallucinate\nphoto-realistic high-resolution (HR) faces from low-resolution (LR) faces, they\ncannot guarantee preserving the identities of hallucinated HR faces, making the\nHR faces poorly recognizable. To address this problem, we propose a Siamese GAN\n(SiGAN) to reconstruct HR faces that visually resemble their corresponding\nidentities. On top of a Siamese network, the proposed SiGAN consists of a pair\nof two identical generators and one discriminator. We incorporate\nreconstruction error and identity label information in the loss function of\nSiGAN in a pairwise manner. By iteratively optimizing the loss functions of the\ngenerator pair and discriminator of SiGAN, we cannot only achieve\nphoto-realistic face reconstruction, but also ensures the reconstructed\ninformation is useful for identity recognition. Experimental results\ndemonstrate that SiGAN significantly outperforms existing face hallucination\nGANs in objective face verification performance, while achieving\nphoto-realistic reconstruction. Moreover, for input LR faces from unknown\nidentities who are not included in training, SiGAN can still do a good job.","url_abs":"http://arxiv.org/abs/1807.08370v1","url_pdf":"http://arxiv.org/pdf/1807.08370v1.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":"sigan-siamese-generative-adversarial-network","repo_url":"https://github.com/jesse1029/SiGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"face-hallucination","task_name":"Face Hallucination"},{"task_slug":"face-reconstruction","task_name":"Face Reconstruction"},{"task_slug":"face-verification","task_name":"Face Verification"},{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"hallucination","task_name":"Hallucination"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.08370","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}