2019/11/20 by Matteo Testa, Arslan Ali, Testa, Matteo +5
Computer Science · #Authorship Attribution and Profiling #Biometric Identification and Security #Computer Vision and Pattern Recognition (cs.CV) #Digital Media Forensic Detection #FOS: Computer and information sciences #Face recognition and analysis #Machine Learning (cs.LG) #Multimedia (cs.MM)
paper · pdf · doi:10.48550/arxiv.1911.08764
openalex publication_date 2019/11/20 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We propose a novel architecture for generic biometric authentication based on\ndeep neural networks: RegNet. Differently from other methods, RegNet learns a\nmapping of the input biometric traits onto a target distribution in a\nwell-behaved space in which users can be separated by means of simple and\ntunable boundaries. More specifically, authorized and unauthorized users are\nmapped onto two different and well behaved Gaussian distributions. The novel\napproach of learning the mapping instead of the boundaries further avoids the\nproblem encountered in typical classifiers for which the learnt boundaries may\nbe complex and difficult to analyze. RegNet achieves high performance in terms\nof security metrics such as Equal Error Rate (EER), False Acceptance Rate (FAR)\nand Genuine Acceptance Rate (GAR). The experiments we conducted on publicly\navailable datasets of face and fingerprint confirm the effectiveness of the\nproposed system.\n