Facial Expression Recognition Using Spatiotemporal Boosted Discriminatory Classifiers

This paper introduces a novel approach to facial expression recognition in video sequences. Low cost contour features are introduced to effectively describe the salient features of the face. Temporalboost is used to build classifiers which allow temporal information to be utilized for more robust recognition. Weak classifiers are formed by assembling edge fragments with chamfer scores. Detection is efficient as weak classifiers are evaluated using an efficient look up to a chamfer image. An ensemble framework is presented with all-pairs binary classifiers. An error correcting support vector machine (SVM) is utilized for final classification. The results of this research is a 6 class classifier (joy, surprise, fear, sadness, anger and disgust ) with recognition results of up to 95%. Extensive experiments on the Cohn-kanade database illustrate that this approach is effective for facial exression analysis.