Improve Face Recognition Using Uncalibrated Photometric Stereo

Abstract

One characteristic of surface image is that the appearance of the surface which is a function of the illuminant direction as well as of the surface topography. Uncalibrated photometric stereo (UPS) is a method used for estimating the surface normal and the surface reflectance of images without a priori knowledge of the light-source direction or the light-source intensity. It is used for shape estimation, surface description and determines illumination direction with several kinds of applications and resources, mainly used for
machine vision techniques.
In this paper a face recognition system using Uncalibrated Photometric Stereo technique was proposed. In this system a face image classified as either known or unknown depending on the surface features that extracted from an image using UPS. The system is trained with known faces images to classifying the newly coming test image into one of the classes is the main
aspect of the classification system.
Many image databases with different types, poses, and illuminations were used. The measurement of the performance, the efficiency, and the robust of this system were presents. The experiments demonstrate that the proposed system gives high classification rate results, even when the lighting condition changes or when the images were noises. Also the face classification system
using the UPS gives better results from the view of classification rates and execution times as compared with other techniques.