A Study on Gaze-Controlled PIN Input with Biometric Data Analysis

Abstract

Common methods for checking a user's identity (e.g., passwords) do not consider personal elements characterizing a subject. In this paper, we present a study on the exploitation of eye information for biometric purposes. Data is acquired when the user enters a PIN (Personal Identification Number) through the gaze, by means of an on-screen virtual numeric keypad. Both identification (i.e., the recognition of a subject in a group) and verification (i.e., the confirmation of an individual's claimed identity) are considered. Using machine learning algorithms, we performed two kinds of analysis, one for the entire PIN sequence and one for each key (i.e., digit) in the series. Overall, the achieved results can be considered satisfying in the context of "soft biometrics", which does not require very high success rates and is meant to be used along with other identification or verification techniques-in our case, the PIN itself-as an additional security level.