Gaze based user authentication for personal computer applications

Abstract

This paper presents a simple PIN-like approach to user authentication, using the gaze sequence of an observer when presented with a previously-seen image on a personal computer screen. The method relies on the principle that the human visual system requires the eye to rest motionless for short periods, to assimilate detail at a given location in a visual scene. By deliberately looking at certain features or objects in a scene following a pre-defined sequence specified by the observer, an independent signature for personal identification can be established. Points of gaze fixation can be identified using a simple eye-tracker based on images from a typical webcam, and processed to extract a compact representation of the significant screen locations from the gaze sequence. Experimental results demonstrate that these "signatures" can be reliably and rapidly computed, and offer the advantage that they are difficult to detect by covert means.