An Efficient Method for Face Feature Extraction Based on Contourlet Transform and Fast Independent Component Analysis

In this paper, an efficient feature extraction method based on the discrete contourlet transform using fast independent component analysis (FastICA) and the angle similarity coefficient(cosine) as the distance measure is proposed. Firstly, each face is decomposed using the contourlet transform. The contourlet coefficients of low and high frequency in different scales and various angles are obtained. The frequency coefficients are used as a feature vector for further processing. Secondly, considering the specificity of face images, we adopt the FastICA algorithm based on negentropy to extract the face feature information. Finally, we according to the distance to classify face feature. Experiments are carried out using the ORL databases. Preliminary experimental results show that the recognition rate and robustness of the proposed algorithm is acceptable and very promising, and confirm the success of the proposed face feature extraction approach.