Facial Expression Recognition

Abstract

Analysis and recognition of facial expression is an important aspect to implement the intelligent man-machine interface. It is also a significant component of artificial intelligence and research of emotion computing. It plays a significant role for illiterate and visually challenged users to access information. Mode of human-computer interaction speech, text, gestures, facial expressions, symbols, or a combination of these. This paper presents a detailed comparative analysis of expression recognition across multiple databases. The system uses Gabor and Log-Gabor filter for feature extraction. Then facial emotion is recognized using both K-Nearest Neighbour KNN and Artificial Neural Network ANN technique. Various test cases are explored in detail using which we conclude that log-Gabor filter and ANN gives best results with 96.4% accuracy, even across multiple databases.