Independent component analysis for face recognition based on two dimension symmetrical image matrix
暂无分享,去创建一个
The new face recognition method based on the matrix of symmetrical face image ICA is put forward for the problem that the influence of the light on the face recognition and high dimensional small sample exists in traditional independent component analysis (ICA)in face recognition. At the same time, in order to improve the human face recognition efficiency, the ICA face recognition method based on the matrix of symmetrical face image is put forward. The method uses the natural characteristics with mirror symmetry of face. According to parity decomposition principle, the odd and even symmetrical samples are created. And symmetrical face image is used as training sample. The principal component analysis (PCA) is used to remove second order relevant and reduce dimension, and then the handled sample is feature extracted by ICA. According to the theory analysis and experimental proof, the influence caused by view, light, face expression, the posture change factors on the face is effective reduced by the new algorithm. Meanwhile, the algorithm increases the size of training sample and reduces the complexity of calculation. In the meantime, the algorithm solves the problem of small sample and improve face recognition rate.
[1] Hagit Hel-Or,et al. Symmetry as a Continuous Feature , 1995, IEEE Trans. Pattern Anal. Mach. Intell..
[2] Aapo Hyvärinen,et al. Fast and robust fixed-point algorithms for independent component analysis , 1999, IEEE Trans. Neural Networks.
[3] Thomas G. Dietterich. Multiple Classifier Systems , 2000, Lecture Notes in Computer Science.