Research And System Development Of EEG-based Emotion Recognition

Emotion plays an import role in various aspects for human life,as the process of judgment and decision making cannot avoid the influence of inner emotions.It has been demonstrated that it is possible to recognize human emotions using electroencephalogram(EEG)signals,so that the EEG based emotion recognition research has great application prospects.In recent few years,the development of machine learning technology has provided reliable techniques for EEG based emotion recognition research.Traditional machine learning methods extract features from multi-channel EEG signals in each channel and concatenate them into a single feature vector,ignoring critical temporal dynamics.The Long Short-Term Memory(LSTM)in deep learning technology can solve this defect well due to the recurrent structure.In this paper,a new emotion recognition method based on LSTM is presented.Various features in time domain,frequency domain and nonlinear dynamic are extracted from multi-channel EEG signals and constructed into feature sequences which are used to train the LSTM based model.Experiments are carried out on the DEAP dataset for valence,arousal and liking classification respectively,and every emotional dimension is divided into two classes(low and high).Experimental results demonstrate that the classification accuracy of the proposed model outperforms the previous methods,with regard to both of valence and liking emotion dimensions,and is also comparable to the most advanced method for arousal classification.What's more,the emotion detection system based on EEG signals is developed.The system mainly contain five modules: central control module,EEG preprocessing module,EEG feature extraction module,emotion detection module and database module.Central control module is responsible for user to choose the operations to be performed.EEG preprocessing module provides many methods for EEG signal analyzing and processing,such as noise reduction and frequency band extraction.EEG feature extraction module is using for extracting features which are relevant to the emotion detection.Emotion detection module can detect subject's emotion change based on the EEG signal.Database module is used to store the information of users,subjects and the detecting results.Users can easily observe the subject's emotion change and statistical analysis on valence,arousal and liking dimension from the client.