Improved Usability for Seamless User Verification based on Biometrics

Abstract

Secure authentication is a crucial service for several systems and applications, particularly when the functionalities offered are extremely confidential or critical. In that situation, single-occurrence, a traditional single authentication mechanism is found to be inadequate for such systems as no verifications are performed after login. In most systems, the authentication is repeated thus results in poor system usability. Continuous biometric authentication is rising as a viable alternative that can promise reliable and transparent verification for the entire session. The presented work focuses on multimodal continuous biometric authentication which incorporates data from facial pictures and keystrokes of the user. The primary intellection behind utilizing more than one biometric is to enhance the recognition rate and to improve usability. The existing systems employ multimodal biometrics consists of physiological and behavioral traits. But they fail in usability as the systems are complex and involve multiple traits. This paper desires to attain a higher security level by adding multiple biometric attributes i.e. face and keystroke dynamics to the continuous user authentication process as both the traits are noninvasive. The proposed system is evaluated and found to be user-friendly and non-invasive when compared with the existing approaches and achieved the Equal Error Rate (EER) of 3.56%.