Abstract
Even after almost three decade of research on automatic face recognition, identification results cannot be considered comparable to superior biometrics. Reasons have been attributed to various modes of variations such as pose, illumination and expression. With the advent of video based face recognition a decade ago we were presented with some new opportunities, algorithms were developed to take advantage of the abundance of data and behavioral aspect of recognition. But this modality introduced some new challenges also, one of them was the variation introduced by speech. In this paper we present a novel method of handling this variation by selecting keyframes from videos based on the temporal analysis of lip motion. Evaluation was carried out by comparing face recognition results obtained by using keyframes selected by the proposed method and frames randomly selected from the videos.