ABSTRACT
A multipose facial expression recognition system has been proposed in this chapter. The proposed system has three components such as (i) image preprocessing, (ii) feature extraction, and (iii) classification. During image preprocessing, from the input body silhouette image, the facial landmarks detection has been performed. Then using these landmark points the rectangular-box face region is detected. Here both frontal and profile face images have been considered for recognizing the type of facial expression in the image. Then a multi-level texture feature descriptors analysis approach is being employed to analyze the texture patterns in the detected facial region for feature computation. For this multi-level feature computation, various global and local feature representation schemes have been employed to extract more distinctive and discriminant features. Finally, a multi-class classification technique using support vector machine classifier is employed to perform classification task to detect the type of facial expression in the image. The performance of the proposed system has been tested on one challenging benchmark facial expression databases such as Karolinska directed emotional faces (KDEF) database and then the performances have been compared with 88the existing state-of-the-art methods to show the superiority of the proposed system.
