3D face reconstruction and recognition using stereo images.

Abstract

Face recognition in still images and extending to video is probably one of the most famous current, past and even futuristic applications and research areas of computer vision and image processing. Relatively a very old history - More than 30 years - of research is involved in this very useful category that is always progressing and enhancing. This work focuses mainly on face recognition using 3D or depth information. We present a fast stereo matching algorithm that can run in real time to generate depth information for a face given its left and right stereo images. The depth information is then augmented with the 2D face information for better recognition against a known labeled and well presented database of face images. Developing fast and accurate face recognition schemes along with the already available fast face detection schemes can lead to a full face or even object recognition system aiding humans in many useful applications.