EEG Emotion Classification Based On Baseline Strategy

Abstract

Electroencephalograph (EEG) emotional computing, as an important task of pattern recognition, has received more and more attention in recent years and is widely used in human-computer interaction, emotional computing and medical fields. Most researches have focused on finding particularly effective features and classifiers to achieve higher classification accuracy in some cases, while most methods are only effective under specific tasks or data and lack broad applicability. In this paper, we propose a novel baseline strategy that introducing emotional features to acquire a newly generated baseline and then, calibrate the individualized features in the emotional features, so that decrease the experimental errors and improve the versatility and effectiveness of the classification method. We performed a classification comparison experiment with baseline-strategy and no baseline-strategy on the DEAP dataset. The selected methods adopt different power spectral density (PSD) feature extraction methods and are classified by Support Vector Machine (SVM) and convolutional neural network (CNN) respectively. The results showed that the experiments with the baseline-strategy achieved better classification results.