Face Recognition Based on a 2D Gabor-Modular Binaryzation-LDA Feature Extraction Method

  • Wenbiao Jiang
  • Mei Xie
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 238)


In this chapter, a novel feature extraction method (Gabor-modular binaryzation linear discriminant analysis (GMBLDA)) for face recognition is presented. 2D Gabor wavelet feature extraction is more robust for illumination and facial expression, however, the dimension of Gabor wavelet is very high and the redundancy rate becomes larger when it is used to feature extracting. For heavy work calculating, the poor real-time, and some other problems, this chapter proposes a way of integrating 2D Gabor-LDA with blocking and binaryzation. Transform face images to 2D Gabor wavelet at first and then use the blocking and binaryzation to reduce the dimensions and finally use LDA transformation to obtain the optimal classification characteristics. Experiment results based on YALE face database demonstrate that this method can achieve a higher recognition rate and a better recognition effect compared with the traditional way such as Gabor, principal component analysis (PCA), LDA, and Gabor + LDA (GLDA).


  1. 1.
    Turk M, Pentland A (1991) Eigenfaces for recognition. J Cogn Neurosci 3(1):71–86CrossRefGoogle Scholar
  2. 2.
    Lee TS (1996) Image representation using 2D Gabor wavelets. IEEE Trans Pattern Anal Mach Intell 18(10):859–970Google Scholar
  3. 3.
    Martinez AM, Kak AC (2001) PCA versus LDA. IEEE Trans Pattern Anal Mach Intell 23(2):228–2332CrossRefGoogle Scholar
  4. 4.
    Pang Yw, Zhang L, Li MJ (2005) A novel Gabor-LDA based face recognition method. Advances in multimedia information. Lect Notes Comput Sci 3331:352–358Google Scholar
  5. 5.
    Tao D, Li X, Wu X et al (2007) General tensor discriminant analysis and Gabor features for gait recognition. IEEE Trans Pattern Anal Mach Intell 29(10):1700–1715CrossRefGoogle Scholar
  6. 6.
    Phillips PJ (2002) Face recognition vendor test evaluation report.

Copyright information

© Springer Science+Business Media New York 2014

Authors and Affiliations

  1. 1.Processing and Information Security LabCollege of Electronic Engineering, UESTCChengduChina

Personalised recommendations