A novel facial expression recognition based on the curevlet features

Abstract

Curvelet transform has been recently proved to be a powerful tool for multi-resolution analysis on images. In this paper we propose a new approach for facial expression recognition based on features extracted via curvelet transform. First curvelet transform is presented and its advantages in image analysis are described. Then the coefficients of curvelet in selected scales and angles are used as features for image analysis. Consequently the Principal Component Analysis (PCA) and Linear Discriminate Analysis (LDA) are used to reduce and optimize the curvelet features. Finally we use the nearest neighbor classifier to recognize the facial expressions based on these features. The experimental results on JAFFE and Cohn Kanade two benchmark databases show that the proposed approach outperforms the PCA and LDA techniques on the original image pixel values as well as its counterparts with the wavelet features.