Emotion Detection Using Relative Grid based Coefficients through Human Facial Expressions

Facial expressions are important in human computer interaction, because the machine can thereby understand human reactions and act accordingly. Facial expressions act like a nonverbal communication cues in human-human or human computer interactions. In noises environment, getting visual data is difficult. According to the relative bin sub-image based studies, high dimensionality is affecting the system which consequently affects the performance of the emotion detection system. Due to these reasons, a new approach using relative grid coefficient feature extraction through visual data is proposed. Support vector machine with radial basis kernel is used for the classification of emotions. Preliminary results showed that an average of 89% accuracy was obtained for relative grid based features.