RECOGNITION OF EMOTION INTENSITIES USING MACHINE LEARNING ALGORITHMS: A COMPARITIVE STUDY

Abstract

A lot ofwork has been done in the recent years for automatic recognition offacial features but the existing work do not measure the intensity offacial emotion. Increasing need for behavioral biometric systems and human-machine interactions demands recognition where intensity ofthe emotions is also measured. Therefore, in this study the aim is to detect various facial emotions (happy, sad, fear, disgust, neutral, surprise and angry) using machine learning algorithms. In this project, HOG and Gabor filter are used for features extractions where as SVM and KNN algorithm are used for emotions classification. The accuracy achieved for SVM is 96.62%. and the accuracy for KNN is 94.62%. The results validate that the facial emotions can be detected with good accuracy using machine learning algorithms. Further this study emotion and intensity of emotion recognition. more efficient facial ^motion can be improved by performing real-time behavioral facial