Facial Expression Recognition Using SVM Classification in Perceptual Color Space

Facial expression analysis is an important area of Human Robot Interaction (HRI) because facial expressions represent human emotions. Here, a new facial expression recognition system is introduced which uses tensor concept. Here perceptual color space is used instead of RGB color space since it cannot work well with illumination and pose variations. Also for classification purpose SVM classifier is used. The experimental results are compared using accuracy and the proposed method shows significant improvement in terms of these factors.