Development of facial expression recognition for training video customer service representatives

This paper describes a study of the relation between facial expression and customer impression of service quality. Based on the results, a facial expression warning system will be designed to improve the service quality of the Customer Service Representative when they practice in training sessions. The system, based on existing systems, has three modules: facial recognition, feature extraction and facial expression recognition. This paper also uses Haar-like features for face detection and a Support Vector Machine for facial expression recognition in an attempt to improve the recognition rate by using a novel feature extraction method. This paper also presents results of the method when tested on standard facial expression databases.