Machine learning blood pressure monitoring method based on LSTM neural network

Abstract

The invention discloses a machine learning blood pressure monitoring method based on an LSTM neural network. The method comprises the following steps: acquiring electrocardio pulse wave signals and pressure pulse wave signal data of a monitored object through a signal acquisition device; carrying out preprocessing of noise removal and baseline drift on the obtained signal, and carrying out filtering to obtain an electrocardiogram wave and a PPW waveform; segmenting and processing the electrocardiogram waveform signal data and the PPW signal waveform signal data; and performing feature extraction on the segmented data to obtain a feature vector sequence, and inputting the feature vector sequence into an LSTM neural network model, wherein the output information of the LSTM neural network isthe blood pressure information of the monitored object and comprises systolic pressure SBP and diastolic pressure DBP. The key parameter PTT can be effectively utilized, and many characteristics of pulse waves of a part of people are difficult to recognize, so that universality of blood pressure monitoring is improved by only recognizing necessary characteristics.

Description

technical field [0001] The invention belongs to the medical field, and in particular relates to a method for continuous blood pressure monitoring, and more specifically designs a method for continuously monitoring systolic blood pressure and diastolic blood pressure by using LSTM neural network to extract multi-feature mixing using ECG and pressure pulse wave signals method. Background technique [0002] Blood pressure is the pressure of blood on the side of the blood vessel wall, which is produced by the blood flowing through the blood vessel when the heart contracts, and is a comprehensive reflection of hemodynamic factors such as circulating blood volume, cardiac output, and arterial wall elasticity. Systolic blood pressure and diastolic blood pressure are important physiological indicators that can reflect the functional status of the human heart and blood vessels. At the same time, they are also one of the four basic characteristics of human health. [0003] The curren...

AI Technical Summary

Problems solved by technology

[0004] The shape of the photoplethysmography (PPG) waveform varies greatly from person to person. For subjects with cardiovascular problems, it is difficult to extract the dicrotic wave waveform of PPG, so only by extracting the characteristics of the PPG waveform has a wide range Limitation, the accuracy will be relatively low

It is convenient to use PTT to calculate blood pressure detection and fitting, but there are also the following problems: (1) The pulse wave of the human body is easily disturbed by multiple factors, including respiration, impedance, measurement posture, etc. Only one parameter of PTT is used to calculate blood pressure accurately low degree

(2) PTT has a higher correlation with systolic blood pressure and a lower correlation with diastolic blood pressure

However, the traditional pressure pulse wave-based measurement method extracts too many features from the waveform, which are susceptible to interference, which affects the measurement accuracy and has a narrow scope of application.

Images