Simple 1D Discrete Hidden Markov Models for face recognition

Abstract

We propose an approach to cope with the problem of 2D face image recognition system by using 1D Discrete Hidden Markov Model (1D-DHMM). The Haar wavelet transform was applied to the image to lessen the dimension of the observation vectors. The system was tested on the facial database obtained from AT&T Laboratories Cambridge (ORL). Five images of each individuals were used for training, while another five images were used for testing and recognition rate was achieved at 100%, while significantly reduced the computational complexity compared to other 2D-HMM, 2D-PHMM based face recognition systems. The experiments done in Matlab took 1.13 second to train the model for each person, and the recognition time was about 0.3 second. © Springer-Verlag Berlin Heidelberg 2003.