Eigenhill vs. eigenface and eigenedge

In this study, we present a new approach to overcome the problems in face r ecognition associated with illumination changes by utilizing the edge image s rather than intensity values. However, using edges directly has its probl ems. To combine the advantages of algorithms based on shading and edges whi le overcoming their drawbacks, we introduced "hills" which are obtained by covering edges with a membrane. Each hill image is then described as a comb ination of most descriptive eigenvectors, called "eigenhills", spanning hil ls space. We compare the recognition performances of eigenface, eigenedge a nd eigenhills methods by considering illumination and orientation changes o n Purdue A & R face database and showed experimentally that our approach ha s the best recognition performance. (C) 2000 Pattern Recognition Society. P ublished by Elsevier Science Ltd. All rights reserved.