Liveness Detection Using Gaze Collinearity

Abstract

This paper presents a liveness detection method based on tracking the gaze of the user of a face recognition system using a single camera. The user is required to follow a visual animation of a moving object on a display screen while his/her gaze is measured. The visual stimulus is designed to direct the gaze of the user to sets of collinear points on the screen. Features based on the measured collinearity of the observed gaze are then used to discriminate between live attempts at responding to this challenge and those conducted by "impostors" holding photographs and attempting to follow the stimulus. An initial set of experiments is reported that indicates the effectiveness of the proposed method in detecting this class of spoofing attacks.