Low complexity 2-D Hidden Markov Model for face recognition

Author

Otluman, H. ; Aboulnasr, Tyseer

Author_Institution

Sch. of Inf. Technol. & Eng., Ottawa Univ., Ont., Canada

Volume

5

fYear

2000

fDate

2000

Firstpage

33

Abstract

In this paper, a low complexity 2-D Hidden Markov Model (HMM) Face Recognition (FR) system is introduced to provide a 2-D representation of the statistical features of the facial image, as opposed to the 1-D HMM and the 2-D Pseudo HMM (2-DPHMM) found in the literature. The proposed model is designed to have low complexity when compared to the Markov Random Field based 2-D HMM (MRF 2-D HMM). The model is implemented in the 2-D Discrete Cosine Transform (DCT) compressed domain based on a non-overlapped 8×8 pixel blocks scheme to maintain the compatibility with JPEG. It is shown that the proposed model has considerably lower complexity than the MRF 2-D KMM and 2-D PHMM

Keywords

Viterbi decoding; computational complexity; discrete cosine transforms; face recognition; hidden Markov models; image representation; 2-D discrete cosine transform compressed domain; 2-D representation; JPEG compatibility; face recognition; facial image statistical features; low complexity 2-D hidden Markov model; modified Viterbi algorithm; nonoverlapped 8×8 pixel blocks scheme; Automatic speech recognition; Discrete cosine transforms; Discrete wavelet transforms; Face recognition; Hidden Markov models; Image recognition; Pixel; Stochastic processes; Strips; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 2000. Proceedings. ISCAS 2000 Geneva. The 2000 IEEE International Symposium on

Conference_Location

Geneva

Print_ISBN

0-7803-5482-6

Type

conf

DOI

10.1109/ISCAS.2000.857356

Filename

857356