Analysis of facial expressions using PCA on half and full faces

Face recognition and expression analysis is one of the most challenging research areas in the field of computer vision. Even though face exhibits different facial expressions, which can be instantly recognized by human eyes, it is very difficult for a computer to extract and use the information content from these expressions. In this paper we present a method to analyze facial expression by focusing on the regions such as eyes, mouth etc whose geometries are mostly affected by variation in facial expressions. Face regions are recognized using principal component analysis (PCA) method. Face images are projected on to a feature space and the weight vectors are compared to get minimum variation. The geometric coordinates of highly expression reflected areas are extracted for analyzing facial expressions. Our method reliably works even with faces, which carry heavy expressions. A comparative study was done by exploiting the symmetrical structure of faces. Our approach performed well for individual half regions of faces. This method exhibits a good performance ratio.