The study of face recognition based on hybrid principal components analysis and independent component analysis

In this paper a new method has been proposed based on the combination of principal component analysis(PCA) and independent component analysis (ICA) for face recognition. In the process of face feature subspace extraction, the dimension of the date set has been reduced by the principal component analysis (PCA), and then the principal components are used to construct eigenface subspace as R. At the same time, the solution of mixed matrix W can be obtained by the fast fixed point algorithm (Fast ICA) iteration, thereby the face feature classification subspace R^TW can be got in the end. In the face classification, the recognition vector can be got when the training face set and testing face set have projected in face feature subspace respectively, then the classification results can be obtained while the vector is brought into the K-neighbor classifier. In this paper the ORL face database has been used to verify this method. The experiment result shows that the algorithm has a good recognition rate.