Emotion Recognition on the Basis of Eye and Mouth

Abstract

It is very interesting to identify human facial expression in the domain of image processing. Since last decade, a lot of research has been focused on this area. The role of this field in human–computer interaction has increased its importance multifold. In this paper, we have used manual database of eyes and mouth which is further classified on the basis of emotion. The expressed emotions included anger, sadness, happiness, and neutral. This classified database is then used to retrieve corresponding facial emotion. We have used Haar cascade technique for detecting the eyes and mouth and nearest neighbor for classification of emotion. The accuracy of this technique is measured on 400+ sample images taken arbitrarily. From the result, we conclude that presented algorithm is able to precisely classify emotion in frontal images on the basis of eyes and mouth only.