THE USE OF INFORMATION CLASSIFICATION IN FACE RECOGNITION AND IDENTIFICATION USING EIGENFACES

Abstract

This research looks into how information classification of face images can be used to improve the efficiency of face recognition and identification system. Two face recognition and identification systems are built; one that does not use information classification, while the other uses information classification. A system without information classification, named FaceR System runs without using additional parameters, whereby the database for this system contains only the face images of a person along with their personal details, such as Name, Id and so on. Whereas a system with information classification, named FaceRpar System runs using additional parameters, in which the database does not only contain the face images and the person’s details, it also extends the person’s details to include other data such as gender and race. In this context, the purpose of the information classification is to classify the face images for the training set based on gender and race. As a result, the size of the face images in the training set decreases after the information classification process. Eigenface algorithm is used for recognizing faces. For both systems, the implementation is tested for both frontal and profile view faces with neutral, speech and smiling expressions of a standard facial image. The performances of the both systems are measured by looking at the processing time and the accuracy of the recognised and identified face. The implication of this method shows that the processing time to execute the face recognition and identification process improves and the identification process becomes more accurate when information classification is used.