Abstract:
Emotion regulation and relaxation state assessment aiming at relaxation and stress reduction can help improve the physical and mental health of the people. In order to reduce the difficulty of inducing relaxation caused by the acquisition of physiological signals, this paper uses the attached human sensor to collect the heart rate signal of the subjects to identify the relaxation state. Mixed audio is used to induce the subjects to produce a relaxed mood, and the relaxed state label is converted from the two-dimensional mood scale. The heart rate signal of the subject is collected, and the time domain feature, frequency domain feature and heart rate are extracted from the heart rate signal. Based on multilayer perceptron and long-short-term memory network, a relaxation evaluation model is constructed to realize relaxation state recognition. The experimental results show that, compared with the current research results, the relaxation recognition model proposed in this paper has better classification performance, and may provide a new and reliable method for emotion regulation and relaxation state assessment.