An analysis of facial expressions from motion history image using PCA and SVM

Real time facial expression recognition approach is proposed in this paper. Since the human face plays a vital role in expressing emotions, the work exploits the Motion History Image (MHI) of face region to extract the motion information. The MHI of the face region is divided into blocks and block based motion history intensity code (BMHI) extracted from the motion information is used as a feature in this work. The extracted feature is further processed using principal component analysis (PCA) to reduce the dimension. The experiments are carried out on real time image sequences, by considering three facial expressions viz., (happy, disgust and surprise) and two strange expressions. Support Vector Machine (SVM) is used to classify the facial expressions. The approach achieves an overall accuracy of 94.13%.