Automatic person recognition using eye movement during scene understanding

Abstract

The human eye is rich in physical and behavioral attributes that can be used for automatic person recognition. Eye movement is a behavioral attribute that has been used for biometric recognition. Usually a task oriented visual stimuli is presented to the subject and his eyes are tracked using a video camera, which are then used as biometric identifier. The most common visual stimuli employed includes the moving object and free viewing. In this paper we have experimented with a novel task oriented visual stimuli i.e. scene understanding. In scene understanding the participants are instructed beforehand that they must perform a task based on the contents of the image/video that will be presented. The information regarding the task adds a considerable cognitive component to the viewing process which in turn impacts the eye movements. We have developed a complete biometric recognition system based on the eye movements extracted during scene understanding based stimuli. Furthermore we have tested the proposed system on several publicly available databases and the preliminary results presented in this paper with an error rate of 17.65 % are quite promising.