DIAGNOSING THE LEARNING DIFFICULTY WITH FACIAL EXPRESSION RECOGNITION

Abstract

Face recognition is extensively applied in this monitoring system and the identity system. In this study, image pre-processing, face detection, and facial expression identification techniques are combined to diagnose learners’ learning difficulty. Face image noise from a WebCam is dealt with integrative type of filter. Then, the feature at the edge is extracted through Sobel operator. After edge enhancement, the eye/mouth images are extracted by skin detection. Three criteria for diagnosing learning difficulty are selected as vectors of Approach-Withdrawal, Concentration, and Happiness to identify difficulty levels. Experimental results show that the correct rate of difficulty recognition reaches 92%. The proposed facial expression recognition system can be used to diagnose students’ learning difficulty in an adaptive e-learning system.