EMOTION RECOGNITION BASED ON EEG FEATURES WITH VARIOUS BRAIN REGIONS

Abstract

Currently researchers have shown immeasurable awareness in Brain Computer Interface (BCI) systems, which enable any user to exchange intelligence and knowledge with surrounding and control instruments by using brain signals; concept is identified as Affective Computing. In this work we are using the SEED database, which is publically available to classify three emotions Positive, Negative and Neutral. Five electrode pairs from various brain regions like Prefrontal, Frontal, Temporal, Parietal and Occipital are selected for this work based on previous research. Diverse time domain and time frequency domain features are extracted from EEG signals. Wavelet Transform (WT) is used to extract a variety of time frequency domain features. Daubechies wavelet function (db6) with 6 levels of decomposition is used to split EEG signals into various frequency bands (δ, θ, α, β and γ). SVM and k-NN algorithms are used as classifiers to estimate classification performance. Hypothetical results illustrate an average classification accuracy of 62.4% for classifying three emotions. Gamma and Beta, the higher frequency bands perform well in emotion recognition.