A Framework for Evaluation of Biometric Based Authentication System
Abhinav Dube, Dhruwaman Singh, Rajesh Kumar Asthana, Gurjit Singh Walia · 2020
In principle, thorough evaluation of performances necessitates the incorporation of a standard well defined suits of criteria and corresponding metrices to analyse biometric based authentication system. Moreover, most of biometric systems are experimentally evaluated over limited metrics and datasets. This not only inhbits to explore the full system potential but also leads to associated risks. Hence, protection of user identity alongwith biometric system from various attacks is paramount for universal acceptance. In order to address this issues, this paper is an attempt to bring all the proposed criteria and methods that targets to evaluate the biometric system for three main parameters-security, privacy and perfomance. These three parameters can define the biometric based recognition system acceptability in an universal manner. To cater the need of more generalized framework for ranking and benchmarkng of the biometric recognition system, a generalized framework has been propsosed. The proposed evaluation framework thoroughly analyses the system based on empirical data collected from all three aspect of evaluation to defines the system acceptability, irrespective of modality and biometric matching algorithms.