Cuff-free blood pressure model generation method, device and equipment and readable storage medium

Abstract

The invention discloses a cuff-free blood pressure model generation method, which comprises the following steps of: initializing parameters of a constructed blood pressure prediction model to obtain ablood pressure prediction meta-model; obtaining training sample data of a plurality of user individuals; performing optimization training based on the training sample data of each user individual andthe blood pressure prediction meta-model to generate an individual model corresponding to each user individual; and optimizing parameters of the blood pressure prediction meta-model according to theindividual model of each user individual, so that the optimized blood pressure prediction meta-model is close to the distance of each individual model in the parameter space, thereby obtaining the optimized blood pressure prediction model. The invention further discloses a cuff-free blood pressure model generation device and equipment and a computer readable storage medium, and the calibration difficulty caused by large model parameter quantity is effectively relieved by constraining the relative positions of the blood pressure prediction meta-model and the individual model.

Description

technical field [0001] The present invention relates to the field of medical devices, in particular to a method, device, equipment and computer-readable storage medium for generating a cuff blood pressure model. Background technique [0002] The current common non-invasive blood pressure measurement methods can be divided into two types: cuff type and cuffless type. Representatives of the cuff method are the Korotkoff method and the oscillometric method, which have the advantage of high accuracy in a single measurement, but because the blood pressure measurement needs to be inflated and deflated by the cuff, continuous monitoring cannot be achieved and it is not convenient for long-term use. Therefore, it is particularly important to develop a cuffless blood pressure measurement method suitable for continuous blood pressure monitoring. [0003] The existing cuffless blood pressure measurement method is greatly affected by individual differences, and needs to be calibrated f...

AI Technical Summary

Problems solved by technology

[0003] The existing cuffless blood pressure measurement method is greatly affected by individual differences, and needs to be calibrated for each individual before use to ensure measurement accuracy

However, the number of calibration samples and calibration time required to calibrate the model increases with the number of model parameters

Especially when deep learning technology is used to model the blood pressure measurement model, the number of model parameters can easily reach tens or even millions. At this time, the number of calibration samples and calibration time required will be very large, which greatly limits the accuracy of cuffless blood pressure. Application of measurement technology

Images