Facial expressions are one of the most powerful, natural and immediate meansfor human being to communicate their emotions and intensions. Recognition offacial expression has many applications including human-computer interaction,cognitive science, human emotion analysis, personality development etc. In thispaper, we propose a new method for the recognition of facial expressions fromsingle image frame that uses combination of appearance and geometric featureswith support vector machines classification. In general, appearance featuresfor the recognition of facial expressions are computed by dividing face regioninto regular grid (holistic representation). But, in this paper we extractedregion specific appearance features by dividing the whole face region intodomain specific local regions. Geometric features are also extracted fromcorresponding domain specific regions. In addition, important local regions aredetermined by using incremental search approach which results in the reductionof feature dimension and improvement in recognition accuracy. The results offacial expressions recognition using features from domain specific regions arealso compared with the results obtained using holistic representation. Theperformance of the proposed facial expression recognition system has beenvalidated on publicly available extended Cohn-Kanade (CK+) facial expressiondata sets.
展开▼
