EEG signal based Modified Kohonen Neural Networks for Classification of Human Mental Emotions

Abstract

Identifying human emotions is very important in human machine interaction(HMI). These emotions will affect communication between people, and their mood.Emotion detection will give a clear idea aboutcustomer satisfaction in e-learning, marketing, entertainments and behavior of criminals in law. Artificial neural networksare essential for machine learning and emotion detection.The emotions are detected from EEG signals which can give better performance to audio and facial signals. In this work, several modified Kohonen neural networks are proposed for human emotion classification.EEG signals from DEAP Database are used asinput for ANN to detect the human emotions. Angry, Happy, Sad and Relax are the emotions classified using KohonenNeural Networks.Experimental results show promising results for the proposed approaches.