Video-based biometric identification using eye tracking technique

Recently, biometric identification techniques have attracted great attention due to increasing demand of high-performance security systems. Compared with conventional identification methods, biometric techniques provide more reliable and robust solutions. In this paper, a novel video-based biometric identification model based on eye tracking technique is proposed. Inspired by visual attention, video clips are designed for subjects to view in order to capture eye tracking data reflecting their physiological and behavioral characteristics. Various visual attention characteristics, including acceleration, geometric, and muscle properties, are extracted from eye gaze data and used as biometric features to identify persons. An algorithm based on mutual information of features is adopted to perform feature evaluation for obtaining a set of the most discriminative features for biometric identification. Experiments are conducted by using two types of classifiers, Back-Propagation (BP) neural network and Support Vector Machine (SVM). Experimental results show that using video-based eye tracking data for biometric identification is feasible. In particular, eye tracking can be used as an additional biometric modal to enhance the performance of current biometric person identification systems.