Non-contact intelligent human body heart rate prediction method based on deep learning

Abstract

The invention discloses a non-contact intelligent human heart rate prediction method based on deep learning, wherein the method comprises the steps: a human face video is acquired through a public data set or a self-shooting video, and ROI extraction of a human face region in the video is performed to acquire a forehead part and a cheek part; then, the video is subjected to fragment division, a time sequence of pixel average values in the ROI changing along with time is obtained through the divided fragments, and the time sequence serves as an initial signal; finally, initial signals are extracted and denoised through multiple signal processing technologies, the signals serve as input of a deep learning model, periodic features related to the heart rate are extracted, and a heart rate result is obtained. According to the method, only a common camera is used for non-contact measurement, expensive and complex equipment such as a thermal infrared camera or a Doppler radar is not needed, the heart rate value is obtained through a signal processing method and a deep learning model, and noise generated by ambient light change and movement is avoided, so that compared with Fourier transform, the obtained heart rate value is more accurate.

Description

Technical field [0001] The present invention relates to the field of intelligent human heart rate prediction techniques, and more particularly to a depth learning-based non-contact intelligent human heart rate prediction method. Background technique [0002] Photoplethysmography, PPG is used to perform non-contact heart rate measurements with PPG, which is mainly based on PPG. PPG is an optical measurement technique that changes the change in blood volume in vascular tissue by analyzing the change in reflected light in the face region. Because the blood absorbs more light than the surrounding tissue, the change in blood volume can correspondingly affect the transmission and reflection of light. [0003] When the skin is irradiated with a beam irradiation, the light beam passes through the reflection, transmission or the like of the skin to the receiving sensor. Human skin tissue, bone, etc., the attenuation of the beam is fixed, but the influence of blood due to cardiac cycle wil...

AI Technical Summary

Problems solved by technology

This method does not extract the features of the heart rate information. Because the ambient light changes and the noise generated by the movement will have a great impact on the final result, the accuracy of the heart rate value obtained by Fourier transform is low.

Images