Využití umělé inteligence pro vícenásobnou bezkontaktní biometrickou autentizaci

Abstract

Biometric authentication systems are used to verify the identity of the person using unique physical features (fingerprint, facial geometry, iris, retina, hand geometry, voice, etc.). The advantage of this type of authentication is that a person does not need to remember a password or always carry an easily stealable token (registration card). Biometric authentication is a fast, convenient and very precise method. Among the main benefits of biometric authentication include a high reliability, zero operating costs, speed, practicality, and clarity. The field of application of biometric systems can be divided into two spheres - security-commercial (security of computers and data, ensuring a comfort, entry into buildings) and forensic (judicial, forensic and investigative). The basic of all biometric systems is automated scanning of biometric characteristics and their subsequent comparison with previously collected data. One of the goal in the field of security is the realization of complex systems based on a combination of multiple characteristics measurements. For this reason, this work focuses on the design of a biometric authentication system based on two characteristics - voice and facial geometry, where the main role in the classification is played by machine learning. By creating such a multimodal autentication system with robustly designed classifiers, a unique contactless authentication tool is created, which can be used as an building access system or personal access system in the future.