Express Recognition Exploring Methods of Emotion Detection

Computer, robotic and mobile interfaces are beginning to use expression recognition to give a more human experience. To make an interface more dynamic and seamless to human interaction, understanding of emotions is key. A robust application to recognize certain facial expressions in real time has many obstacles starting from correctly identifying a face and extracting necessary features of the face to then mapping these features to the right expression. There has been a thorough study of how to extract key points on faces and OpenCV even provides this code freely [1]. A topic that has been over looked is how to best manipulate a minimal number of facial key points efficiently or even if facial key points are the fastest route to facial expression identification. The goal will be to compare different methods of extracting expressions and gain more intuition about this system to further improve how expressions are defined and modeled.