Neural network based classification of human emotions using Electromyogram signals

Facial expression of emotion is of great interest to many researchers. Facial Electromyography (FEMG) is used for the identification of different facial expressions namely happy, sad, fear, neutral, surprise etc. In this paper, a simple algorithm to identify six emotions using the FEMG signals is proposed. FEMG signals are recorded from seven subjects. The six emotions are identified using bandpower features extracted from the raw FEMG signals and neural networks. In this study, two networks are used to identify the emotions. The network has an average classification accuracy of 94.32%.