Simulation study on the performance of several classifiers in face recognition

Abstract

Classifier is the important content in the face recognition system. This paper uses PCA to extract the face image feature and then compares the recognition performance of several classifiers. Based on ORL and YALE face database, the paper carries out simulation experiment by using minimum distance classifier, nearest-neighbor classifier and K-neighbor classifier respectively. Besides, this paper researches different distance measures' effects on the recognition result of the classifiers, Euclidean distance, Minkowski distance, cosine distance and absolute distance.