Classification Of EEG Signal Based On Emotion

The use of computer technology for emotion recognition is the key to achieving high-level human-computer interaction, and it is important for human-computer interaction as well as human-computer interfaces. At the same time, emotion recognition based on physiological signals is more objective. That is why emotion recognition based on physiological signals is one of the hot research topics in the field of pattern recognition, cognitive science, and psychology.Aiming at the research of emotion recognition and classification based on EEG signals, an emotion information collection system is set up firstly, then stimulus files used to induce various emotions, by taking the happy and unhappy as an example, are designed and implemented; the emotion feature extraction based on the principle composition analysis (PCA) and the emotion classification algorithm based on the support vector machine (SVM) are investigated thirdly, and an emotion classification system for happy and unhappy emotion is set up at last. Main works are as follows:(1) A multi-physiological information collection system with image collection subsystem, eye-track information collection subsystem and EEG collection subsystem is designed and realized, in which EEG collection subsystem is focused on in this paper. (2) Based on the distinguishing feature of our university, the picture-stimulating files are researched and designed by combining questionnaire and statistics method, used to induce various emotions with universal significance without losing the national characteristics, in which happy and unhappy picture-stimulating files are focused on in this paper.(3) The mathematical principle of STFT (Short Time Fourier Transform) method is analyzed deeply, and STFT is applied to the EEG signal feature extraction. Under such environment of applying STFT to the he EEG signal feature extraction, how to select window function and the window length is studied comparatively. Hamming window which completely covers instantaneous data after stimulating is the answer obtained from the experiments.(4) PCA algorithm and SVM classification algorithm, especially selecting the kernel function, are studied deeply. Based on the combination of the PCA and the SVM, the ethnic happy and unhappy emotion recognition system is set up. The experiment results demonstrate that the ethnic emotion recognition system with high speed of feature extraction and high average recognition rate is feasible and effective.