Facial Expression Recognition System Using Extreme Learning Machine

Interest is growing in improving all aspect of the interaction between human and computer including human emotions. It is a crucial task for a computer to understand human emotions. A very meaningful way of expressing human emotions is facial expression. In this paper, a model facial expression recognition based on Extreme Learning Machine is proposed. Salient facial feature segments like eyebrows, eyes, mouth, and nose are detected from a face image and then these feature segments are extracted by using morphological image processing operation and edge detection technique to form feature vectors. Extreme Learning Machine, a feed-forward neural network classifier with a single layer of hidden nodes is used for recognizing expressions of the input faces into six basic categories like happy, sad, surprise, angry, disgust, and fear. The experiments of facial expression recognition system are carried out on JAFFE facial expression database and performances of experimental results are analysed.