Adaptive system to learn and recognize emotional state of mind

Facial expression is the most natural way for the humans to convey their emotions with intensions. Facial expressions can be easily detected by human but for machines, it is very challenging task. In this paper, an approach for real time emotion detection is presented using real time graphic user interface for detection of emotion through a web camera. The proposed system named as Adaptive System For Recognizing State Of Mind (ASFRSOM) system is designed using the three major components: face detection, feature extraction and classification. For facial feature extraction, local binary pattern is used and the extracted features are applied to support vector machine for emotion recognition. The approach proposed in the presented work is evaluated on extended Cohn Kanade database and results in 79% accuracy. To further prove the accuracy of the proposed approach, it is used on 15 subjects in real time and emotions were accurately detected.