In this paper, ICA is presented as an efficient feature extraction algorithm used in automatic face recognition task. In a task such as face recognition, important information may be contained in the high-order relationship among pixels. ICA is sensitive to high-order statistic in the data and finds not-necessarily orthogonal bases, so it may better identify and reconstruct high-dimensional face image data. ICA algorithms are time-consuming and sometimes converge difficultly. So a modified FastICA algorithm is developed in this paper, which only need to computer Jacobian matrix once time in one iteration and achieves the correspondent effect of Fast-ICA. After obtaining all independent components, a genetic algorithm is introduced to select optimal independent components (ICs). In this paper, ICA is compared with principle component analysis (PCA) based feature extraction method. The experiment results show that modified FastICA algorithm fast convergence speed and genetic algorithm optimize recognition performance. ICA based features extraction method is robust to variations and promising for face recognition.
[1]
David J. Kriegman,et al.
Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection
,
1996,
ECCV.
[2]
A. O'Toole,et al.
Structural aspects of face recognition and the other-race effect
,
1994,
Memory & cognition.
[3]
Te-Won Lee,et al.
Independent Component Analysis
,
1998,
Springer US.
[4]
David J. Kriegman,et al.
Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection
,
1996,
ECCV.
[5]
Marian Stewart Bartlett,et al.
Face recognition by independent component analysis
,
2002,
IEEE Trans. Neural Networks.
[6]
Aapo Hyvärinen,et al.
Fast and robust fixed-point algorithms for independent component analysis
,
1999,
IEEE Trans. Neural Networks.
[7]
Rama Chellappa,et al.
Human and machine recognition of faces: a survey
,
1995,
Proc. IEEE.
[8]
Andrzej Cichocki,et al.
A New Learning Algorithm for Blind Signal Separation
,
1995,
NIPS.