Parametric Linear Subspace Method for Face Recognition and Pose Estimation

Abstract

A robust face recognition system is described in this paper which is insensitive to human head pose. The combination of Gabor Wavelet Network (GWN) and Parametric Linear Subspace (PLS) method is used for effective and efficient face recognition process. Multiple feature extraction is done and various steps of pose estimation carried out using new algorithms. In this work, the complete algorithm has been carried out in various stages. Initially, source images are applied and flag is set. In the next step, image is converted to gray scale image and the given image's extension is converted into portable gray map. calculate the analysis mapping that is estimate the pose of head from given image, given image is compared with the database, decision is made according to the matched or unmatched templates, above steps are repeated until matching of last database template.