Face Detection using Principal Component Discriminant Analysis

Abstract

Face recognition is an active and important research from past many years. This process includes face tracking, expression finding and many more. In this process it is very important to first register the locations of images. But it is very challenging to maintain such a database because of illumination invariance, pose invariance, noise invariance, shift invariance and scale invariance. By knowing person’s identity, person may be authenticated to utilize a particular service or not.  The Image registration algorithm will register all these images present in the database. This registration algorithms attempt to align a pattern image over a reference image so that pixels present in both images are in the same location.  Therefore the Fourier Mellin [13] Transform is used and it is quite useful for image recognition because its resulting spectrum is invariant in rotation, translation and scale. Taking a pattern classification approach, each pixel in an image can be considered a coordinate in high-dimensional space. Rather than this deviation, project the image into a subspace which discounts the large deviation regions of the face. This is achieved by using dimension reduction techniques like Principal component analysis (PCA), Linear Discriminant analysis (LDA), Laplacian faces and other modified approaches like A Priori Laplacian and PCDA.