{"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/tv-gan-generative-adversarial-network-based","title":"TV-GAN: Generative Adversarial Network Based Thermal to Visible Face Recognition","arxiv_id":"1712.02514","date":"2017-12-07","proceeding":null,"authors":["Teng Zhang","Arnold Wiliem","Siqi Yang","Brian C. Lovell"],"abstract":"This work tackles the face recognition task on images captured using thermal\ncamera sensors which can operate in the non-light environment. While it can\ngreatly increase the scope and benefits of the current security surveillance\nsystems, performing such a task using thermal images is a challenging problem\ncompared to face recognition task in the Visible Light Domain (VLD). This is\npartly due to the much smaller amount number of thermal imagery data collected\ncompared to the VLD data. Unfortunately, direct application of the existing\nvery strong face recognition models trained using VLD data into the thermal\nimagery data will not produce a satisfactory performance. This is due to the\nexistence of the domain gap between the thermal and VLD images. To this end, we\npropose a Thermal-to-Visible Generative Adversarial Network (TV-GAN) that is\nable to transform thermal face images into their corresponding VLD images\nwhilst maintaining identity information which is sufficient enough for the\nexisting VLD face recognition models to perform recognition. Some examples are\npresented in Figure 1. Unlike the previous methods, our proposed TV-GAN uses an\nexplicit closed-set face recognition loss to regularize the discriminator\nnetwork training. This information will then be conveyed into the generator\nnetwork in the forms of gradient loss. In the experiment, we show that by using\nthis additional explicit regularization for the discriminator network, the\nTV-GAN is able to preserve more identity information when translating a thermal\nimage of a person which is not seen before by the TV-GAN.","url_abs":"http://arxiv.org/abs/1712.02514v1","url_pdf":"http://arxiv.org/pdf/1712.02514v1.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":"tv-gan-generative-adversarial-network-based","repo_url":"https://github.com/zachzhu2016/Spectraface","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"tv-gan-generative-adversarial-network-based","repo_url":"https://github.com/zachzhu2016/thermal-face-recognition","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"face-recognition","task_name":"Face Recognition"},{"task_slug":null,"task_name":"Generative Adversarial Network"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}