Physiological parameter measurement method based on multi-scale fusion network

Abstract

The invention provides a method for measuring physiological parameters based on a multi-scale fusion network. The method includes: sampling the physiological signal and generating a one-dimensional physiological signal data sequence; performing noise filtering on the data segment meeting the signal quality requirements after dividing the data segment and evaluating the signal quality; and then performing mathematical transformation on the data segment to generate a multi-dimensional The input data tensor; the multi-scale fusion network is used to extract potential features from the input data to obtain the estimated value of physiological parameters. By identifying the measurement mode identifier, the mean value of all estimated values ​​is used as the measured value of the physiological parameter in the static mode; the one-dimensional continuous data sequence formed by all estimated values ​​is used as the continuous measured value of the physiological parameter in the dynamic mode. This method can fully extract complementary information of different scales in the signal, and realize accurate measurement of physiological parameters. Its application range covers the measurement of all physiological parameters, and has certain application value in the fields of cardiovascular disease research and signal processing research.

Description

Technical field [0001] The present invention relates to a physiological parameter measurement method based on a multi-scale fusion network. Background technique [0002] The heart is the power center of the human blood circulation. Through the regular pulsation to the whole body blood to satisfy the human body's metabolism, thereby maintaining normal life activities. Effectively measuring the physiological parameters associated with cardiovascular system to monitor human health to prevent cardiovascular diseases. With the aging aging and the growth of work pressure, the prevalence of domestic cardiovascular disease is growing year by year. According to the 2020 report of the World Health Organization, cardiovascular disease has become the main cause of human death. Therefore, dynamically detect the physiological parameters associated with cardiovascular system and accurately assess health conditions, which has a major practical significance for the prevention and treatment of car...

AI Technical Summary

Problems solved by technology

At present, most studies use a single-scale convolutional neural network to achieve automatic feature extraction, but this approach ignores potentially useful information on other scales, resulting in the inability to further improve the measurement accuracy

At the same time, for the same physiological signal, the inter-individual variability and the variability of the waveform also make it impossible for a single-scale convolutional neural network to effectively extract features.

Images