Emotion Recognition Based On Cortical Electrical Signals Before And After Exercise

With the rapid development of science and technology,people and computers communicate for longer and longer,and how to make computers capture the real emotional changes of people has become a hot topic.Therefore,affective computing has become an important research direction in human-computer interaction.With the increase of wearable devices,it has become convenient to obtain the user’s physiological signals,because the physiological signals are not interfered by other external factors,and the real and accurate emotional changes of people can be measured.Therefore,more and more scholars use physiological signals to conduct emotion recognition research.Emotion recognition from physiological signals is an important research direction.Since there is currently very little research on the effect of exercise on emotion.Therefore,this article will study the effect of exercise on mood from the changes of skin electrical signals before and after exercise.The main research work is as follows:1.In this thesis,a cortical electric signal emotion dataset is established.First,the emotional videos were manually screened and scored,and a video motivation library was established.In two scenarios: scene 1,before exercise,and scenario 2,after exercise,the filtered video incentive library was used in turn to induce the emotions of the participants.Then,the skin electrophysiological signals of the participants were collected,and the Chinese simplified version of the PAD emotion scale was used for labeling,and finally the better data were selected according to the PAD scale filled in by the participants,and a total of 1400 signals were obtained.2.The collected picoelectric signals are denoised and feature extracted,and then the GSR signals in the two physical states are classified and recognized by SVM,LSTM and ACRNN.The final results show that the recognition effect of happy after exercise is better than that of happy before exercise;After exercise,emotional recognition of sadness and fear was lower than pre-exercise.It shows that people will have an impact on physiological signals after exercise,and the recognition effect of positive emotions is better than that of positive emotions before exercise,and has a certain inhibitory and improving effect on negative emotions.Exercise has a certain impact on people’s physical emotions.3.On the basis of pre-and post-exercise,it was further studied whether gender has an impact on emotions.The collected physiological signal data was divided into two groups,male and female,and then the data were denoised,preprocessed and feature extracted,and input into SVM,LSTM and ACRNN models for emotion recognition.The results showed that the recognition effect of fear and sadness in girls before exercise was higher than that of boys,and the recognition effect of anger and joy in boys was better than that of girls.After exercise,the recognition rate of positive emotions in boys and girls increased,and the recognition rate of negative emotions decreased.But girls have a higher recognition rate of fear and sadness than boys,and happy recognition rate is lower than that of boys.It shows that different genders have different effects on emotions.Figure [ 22 ] table [ 6 ] reference [ 76 ]...