Facial Expression Recognition from Global and a Combination of Local Features

In this paper, facial expressions are identified using contour information, along with the Gabor features extracted from highly expression-reflected facial features. We propose a novel technique to extract Gabor features from a selected region of interests. The computational complexity is reduced by a suitable selection of Gabor filters for the representation of transformed face images. Support Vector Machine classification technique is used to identify the expression that belongs to the face images.