Distance function for face recognition based on 2D PCA

Face recognition based PCA have been attracted extreme interests and thoroughly research since it was first introduced by Turk and Pentland in 1991. In this paper, we discuss the different distance function has play different influence to the result of face recognition based on 2D PCA and sorting to the distance function with parameters according to face recognition efficiency. The rate of recognition is depending on the choice of threshold and threshold will be different when we use different distance function on one set of face image. So we introduce the standard deviation of distance function and sorting distance function directly without choosing the threshold. This result can be applied to general measurement of distance function in face recognition.