End-to-end network-based non-contact heart rate measurement method, system and device

Abstract

The invention belongs to the field of data identification, and specifically relates to a non-contact heart rate measurement method, system and device based on an end-to-end network, aiming to solve how to measure non-contact heart rate without contact and improve the robustness and performance of the measurement method. The problem of generalization ability. The present invention includes: acquiring a sequence of video frames to be tested containing face information, based on the sequence of video frames to be tested, obtaining an image sequence of a region of interest in a face through a face detection model, and obtaining key points of a face through a key point model of a human face The position sequence is based on the face ROI image sequence, and the average heart rate value within a time period is obtained through a trained end-to-end dual-branch network. The invention improves the robustness and generalization ability of the non-contact heart rate detection method, expands the range of heart rate that can be accurately measured, and improves the accuracy of heart rate measurement.

Description

technical field [0001] The invention belongs to the field of data identification, and in particular relates to a non-contact heart rate measurement method, system and device based on an end-to-end network. Background technique [0002] Heart rate is an important indicator that needs to be measured on many occasions, especially in the field of health. In traditional devices, monitoring heart rate and heart activity is accomplished by measuring electrophysiological signals with an electrocardiogram (ECG) or photoplethysmography (PPG), which requires electrodes attached to the body to measure heart tissue. The signal of the electrical activity caused by PPG can only be measured by contacting the device with the human body. The contact device makes the person being tested feel uncomfortable, especially for the surge in demand for non-contact medical care during the epidemic. In order to solve this problem, recent In recent years, the technology of measuring heart rate through f...

AI Technical Summary

Problems solved by technology

The existing methods of testing heart rate through faces still have the following shortcomings: First, the existing data set is not large enough, so we can only use a shallow neural network, and the learned model is relatively easy to overfit, and generally speaking, In the case of sufficient training data, the deeper neural network has stronger discrimination and generalization capabilities; secondly, the task of measuring heart rate through the face is easily affected by head movement, light changes and different equipment; again different losses The function also has a great influence on the result

Images