Contourlets for facial expression analysis using one nearest neighbor classifier

Facial expression is a significant key for interpersonal relations where non verbal cues are conveyed from person to person. Automated recognition of facial expression plays an important role in widely of applications such as human-computer interaction, data-driven animation, clinical practice and behavioral science. In this study a novel and efficient automated recognition of facial expression system is proposed based on Contourlet Transform and k nearest neighbor (KNN) classifier. There are two types of filter banks are used for contourlet construction, which are Non Subsampled Pyramid Structure (NSPF) and Non Subsampled Directional Filters (NSDF). The energy features are extracted from each sub-band of contourlet decomposed image and then recognition is processed by KNN classifier. The performance of the proposed recognition of facial expression system is carried out with two filter banks and the results are compared in terms of classification accuracy. The proposed system attained satisfactory.