The integration of PCA, FLD and gabor two-dimensional wavelet transformation using facial expression feature extraction

In order to improve the rate of facial expression recognition. The Gabor wavelet fusion PCA+FLD method is proposed as a new method this paper. Firstly, face image preprocessing, and then carries on the two-dimensional wavelet transform Gabor, first constructed five scale eight directional wavelet filter, through the PCA+FLD method for dimensionality reduction, and finally obtaining an optimum expression a projection subspace classifier in all the optimal subspace. The experiment proves, fusion PCA, FLD and Gabor wavelet transform two-dimensional facial expression feature extraction for classifier design, facial expression recognition rate is high. The experiment proves that the new method of Gabor wavelet fusion PCA+FLD is perfect.