Non-invasive blood lipid detection method and system based on multi-wavelength ppg signal

Abstract

The invention discloses a non-invasive blood lipid detection method and system based on multi-wavelength PPG signals, comprising: collecting multi-wavelength and multi-channel PPG signals and corresponding invasive blood lipid values ​​of the fingertip of a target object, and denoising the multi-channel signals Process and divide the data set, build a fusion neural network model through convolutional neural network and long short-term memory artificial neural network, train the fusion neural network through the data set, and detect blood lipid content through the trained fusion neural network model, and the detection is accurate. Degree evaluation. The invention uses multi-wavelength light to simultaneously collect PPG signals of different fingertips of the human body, effectively avoids the adverse effect of mutual interference between multi-wavelength light sources on blood lipid evaluation, and at the same time can obtain the time difference of important feature points between signals according to the signal collection of different parts , so as to obtain the relevant information of the blood flow velocity of the human body, and at the same time, the fusion neural network is introduced to improve the accuracy of blood lipid estimation.

Description

technical field [0001] The invention relates to the technical field of blood lipid detection, and more particularly, to a non-invasive blood lipid detection method and system based on multi-wavelength PPG signals. Background technique [0002] The detection of human blood lipids is very important to human health. Abnormal blood lipids in the human body often means that the internal environment of the human body is unbalanced. Hyperlipidemia can cause symptoms such as dizziness, obesity, chest tightness, shortness of breath, and limb numbness. If blood lipids are higher than the normal range for a long time, it will not only cause various diseases such as cardiovascular disease, hypertension, pancreatitis, but even diabetes and atherosclerosis. Or form blood clots, which can seriously affect your health. [0003] According to relevant studies, human blood lipid-related molecules can resonate with light of specific wavelengths. Because the real blood conditions of different ...

AI Technical Summary

Problems solved by technology

[0004] Now, the accurate measurement of blood lipids still needs invasive measurement guarantee, but the cost of invasive measurement is high, and professionals are required to collect samples invasively, and at the same time, there is a high risk of invasive infection; in addition, focusing on different wavelengths of light will affect the surrounding light environment. The requirements are higher, and the accuracy requirements of the equipment are also higher

And the feed-forward neural network in the prior art will think that the input content and results at the next moment are completely irrelevant to the input content and results at the current moment

But in fact, a person's blood lipid value changes regularly over time, and there will be no random mutation increase or decrease, and the network of these prediction models cannot avoid the error of the sudden increase or decrease of the predicted value It is not easy to fix the collection site with the back of the hand, and it is difficult to collect the site according to the positioning signal of the vein blood vessels on the back of the hand

Images