Illumination and noise tolerant face recognition based on eigen-phase correlation filter modified by Mexican hat wavelet

The performance of frequency domain correlation filter is studied for face recognition by considering only the phase part of the face images. The dimensions of the phase spectra are reduced by using frequency domain principal component analysis (PCA). The phase only correlation filter is modified by using continuous Mexican hat wavelet to generate a new wavelet modified phase only correlation filter (WPoCF).The performances of the technique at natural and noisy conditions are improved, when compared to conventional correlation filters. It has been shown that the new approach improves the recognition accuracy at different illuminations and noise conditions.

[1]  B. V. K. Vijaya Kumar,et al.  Correlation Pattern Recognition , 2002 .

[2]  Y Sheng,et al.  Optical wavelet matched filter. , 1994, Applied optics.

[3]  B. V. Vijaya Kumar,et al.  Unconstrained correlation filters. , 1994, Applied optics.

[4]  Stéphane Mallat,et al.  Wavelets for a vision , 1996, Proc. IEEE.

[5]  Rama Chellappa,et al.  Human and machine recognition of faces: a survey , 1995, Proc. IEEE.

[6]  D. Casasent,et al.  Minimum average correlation energy filters. , 1987, Applied optics.

[7]  Tsuhan Chen,et al.  Principle component analysis and its variants for biometrics , 2002, Proceedings. International Conference on Image Processing.

[8]  M. Turk,et al.  Eigenfaces for Recognition , 1991, Journal of Cognitive Neuroscience.

[9]  Andrea F. Abate,et al.  2D and 3D face recognition: A survey , 2007, Pattern Recognit. Lett..

[10]  Pradeep K. Khosla,et al.  "Corefaces" - robust shift invariant PCA based correlation filter for illumination tolerant face recognition , 2004, CVPR 2004.

[11]  Azriel Rosenfeld,et al.  Face recognition: A literature survey , 2003, CSUR.