Abstract: Expression is the most important mode of non-verbal communication between people. Facial expression carries significant information about the mental, emotional and even physical states of the conversation. Recognition of facial expressions has used in wide application areas like scientific, surveillance, medical, marketing etc. This paper proposed hybrid technique is called score level fusion. Initially preprocessing operations performed on the given image and the features are extracted. The proposed work consists of two important phases such as feature extraction phase and fusion phase. Shape features are extracted from the eyes and texture features are extracted from nose and mouth. The eye features are extracted by using Active Appearance Model (AAM) and the nose and mouth features are extracted by using Gray Level Co-occurrence Matrix (GLCM). With the help of the Artificial Neural Network (ANN) classifier one score value is calculated from extracted eye feature and in another case with the help of Adaptive Genetic Fuzzy System (AGFS) score value is calculated from the mouth and nose. The score values are given as input to the fusion phase in which simple-sum fusion method is used for the calculation of the final score. This final score is compared with a threshold value in order to classify the facial expressions.

Keywords: Facial Expressions, Active Appearance Model, Gray Level Co-occurrence Matrix, score value.