A new technique for Face Recognition using 2D-Gabor Wavelet Transform with 2D-Hidden Markov Model approach

Author

Srinivasan, M. ; Ravichandran, Naveen

Author_Institution

Dept. of Electron. & Commun. Eng., Alpha Coll. of Eng., Chennai, India

fYear

2013

fDate

7-8 Feb. 2013

Firstpage

151

Lastpage

156

Abstract

A Discrete Gabor Wavelet Transform (DGWT) based 2D Hidden Markov Model (2DHMM) approach for Face Recognition (FR) is proposed in this paper. To improve the accuracy of the face recognition algorithm, a Gabor Wavelet Transform is used in obtaining the observation sequence vectors. We have conducted extensive experiments ORL database which shows that the proposed method can improve the accuracy significantly, especially when the face image dataset is large with limited training images. Unlike the pervious HMMs used for FR, we propose 2D HMM with Expectation-Maximization (EM)algorithm suitable for almost perfect estimation as feature vectors. This model of 2D HMM shows superior image segmentation for learning process. A recognition rate of 99% is achieved.

Keywords

expectation-maximisation algorithm; face recognition; hidden Markov models; image segmentation; learning (artificial intelligence); vectors; visual databases; wavelet transforms; 2D Gabor wavelet transform; 2D hidden Markov model; 2DHMM; DGWT; EM algorithm; ORL database; discrete Gabor wavelet transform; expectation maximization; face image dataset; face recognition; image segmentation; learning process; observation sequence vector; Face recognition; Hidden Markov models; Vectors; 2D Hidden Markov Model (HMM); Discrete Gabor Wavelet Transform (DGWT); Expectation-Maximization (EM); Face Recognition (FR);

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing Image Processing & Pattern Recognition (ICSIPR), 2013 International Conference on

Conference_Location

Coimbatore

Print_ISBN

978-1-4673-4861-4

Type

conf

DOI

10.1109/ICSIPR.2013.6497977

Filename

6497977