Analysis of Positive and Negative Emotions based on EEG Signal

摘要

The use of computer technology for emotion recognition is the key to achieve high-level human-computer interaction. Aiming at the research of emotion recognition and classification based on EEG signals, emotional stimulus was designed with positive and negative emotions. The emotion feature extraction based on algorithm of Common spatial pattern (CSP) after signal denoising. The emotion classification algorithm based on the support vector machine (SVM), and the finally classification accuracy can achieve 92%. Lastly, we analyze the complexity of EEG data in the by using the wavelet entropy algorithm. It found that wavelet entropy value of the negative emotion state is lower than the positive state. It shows that the brain is nervous and regularity of the brain is stronger when it is negative. In the positive mood, however, the brain is relatively relaxed and the regularity becomes weak. The study realizes the visualization of human emotion, which provides great value for the study of depression and those who can be easily influenced by emotion for heavy stress.