An emotional analysis method based on heart rate variability

In this paper, we propose a method to evaluate human´s emotion and stress based on heart rate variability (HRV). Firstly, experiment scheme has been designed to induce 4 kinds of emotions and the corresponding electrocardiogram (ECG) changes have been measured in a laboratory setting; Secondly, an improved fast denoising method based on wavelet transform threshold denoising was proposed to process the noisy ECG signal, then we realize automatic extraction of HRV sequences; Finally, a wide range of physiological features from various analysis domains, including time, frequency, nonlinear analysis is proposed in order to find the best emotion-relevant features and to correlate them with emotional states, some important conclusions have been obtained.