Research on Signal Enhancement Method in the Measurement of Human Physiological Parameters Based on iPPG

Abstract

The imaging photoplethysmography (iPPG) signal contains rich human physiological information, and can be used to realize non-contact physiological parameter monitoring. However, the signal is weak and susceptible to various factors. At present, the research on the influencing factors of iPPG signal is still in its infancy. Research on the enhancement method of iPPG signal can effectively improve the availability of the acquisition module, thereby improving the accuracy of the physiological parameters obtained. This article studies this problem from threeperspectives: signal acquisition methods, external influences, and signal processing.In the acquisition of the original iPPG signal, the method of enhancing the iPPG signal is to optimize and improve the video face detection and face tracking. This paper uses the compatible method of the Camshift tracking algorithm and the AdaBoost face detection algorithm to improve the ability to detect video faces, and uses an improved Camshift tracking algorithm - ERC (Environmental Robust Camshift) for face tracking and the division of ROI area.Before signal acquisition, the influence of the IPPG signal mainly lies in the light conditions and human body characteristics. Five important influencing factors are mainly studied: light intensity, light color, skin color, age and gender. This paper calculates and compares the SNR of each iPPG signal by using the controlled variable method and Bland-Altman analysis. and the experimental results are only obtained from the signal itself. The results show that light intensity and age have no effect on the SNR of iPPG signals; Green light has the best SNR, followed by white light, blue light, and red light. Bland-Altman analysis does not have consistent data at 5.2%, 7.3%, 8.9%; The average SNR of male volunteers is 0.823, and that of females is 0.784. Males are generally higher than females, but not absolutely; The second, third, and fourth skin colors of men are 1.151, 0.765, 0.541, and women are 0.812, 0.623, 0.318, which means that as the skin colored deepens, the SNR decreases.Signal processing is performed after acquisition, and the method of enhancing the iPPG signal will be carried out on the computer. This article studies two methods of wavelet transform and Euler video amplification technology to enhance the signal. The Euler video amplification technology used in this article starts with the image, which amplifies the frequency of the required signal, and the signal enhancement is naturally realized.