US-12622646-B2 - Method of extracting representative waveform of bio-signal and apparatus for estimating bio-information

US12622646B2US 12622646 B2US12622646 B2US 12622646B2US-12622646-B2

Abstract

A method of extracting a representative waveform of a bio-signal includes receiving an input of the bio-signal; dividing the bio-signal into a plurality of sub-signals; selecting at least one sub-signal for extracting a representative waveform from among the divided sub-signals; extracting a representative waveform by using the at least one selected sub-signal; evaluating a quality of the extracted representative waveform; and based on the representative waveform satisfying a predetermined quality criterion corresponding to the evaluation, determining that the representative waveform is a final representative waveform.

Inventors

Ui Kun Kwon
Young Soo Kim
Chang Soon Park
Dae Geun Jang

Assignees

SAMSUNG ELECTRONICS CO., LTD.

Dates

Publication Date: 20260512
Application Date: 20220215
Priority Date: 20211028

Claims (20)

1 . A method of extracting a representative waveform of a bio-signal, the method comprising: measuring, using a sensor, the bio-signal of an object; dividing the bio-signal into a plurality of sub-signals; selecting at least one sub-signal for extracting a target waveform from among the plurality of sub-signals; extracting the target waveform from the bio-signal by using the at least one selected sub-signal; evaluating a quality of the target waveform; based on the target waveform satisfying a predetermined quality criterion corresponding to the evaluation, determining that the target waveform is the representative waveform; and estimating bio-information of the object based on the representative waveform, the bio-information comprising a blood pressure, an arrhythmia, a vascular age, an arterial stiffness, an aortic pressure waveform, and a fatigue level, wherein the extracting of the target waveform comprises setting first weights of first sub-signals from the at least one selected sub-signal having a first quality that meets or exceeds a predetermined quality threshold based on a difference between the first quality and the predetermined quality threshold, and clearing second weights of second sub-signals from the at least one selected sub-signal having a second quality that does not meet the predetermined quality threshold.
2 . The method of claim 1 , wherein the dividing of the bio-signal into the plurality of sub-signals comprises dividing the bio-signal into the plurality of sub-signals based on units of beats.
3 . The method of claim 1 , wherein the selecting of the at least one sub-signal comprises selecting a predetermined number of sub-signals or sub-signals in a predetermined time interval.
4 . The method of claim 1 , wherein the selecting of the at least on sub-signal comprises selecting a plurality of selected sub-signals; wherein the extracting of the target waveform comprises: setting a reference point in each of the plurality of selected sub-signals; and overlapping the plurality of selected sub-signals based on the reference point in each of the plurality of selected sub-signals.
5 . The method of claim 4 , wherein the setting of the reference point comprises setting, as the reference point, at least one of a minimum point of a waveform of each of the plurality of selected sub-signals, a maximum point of the waveform of each of the plurality of selected sub-signals, a maximum slope point of the waveform of each of the plurality of selected sub-signals, or a tangent intersection point of the waveform of each of the plurality of selected sub-signals.
6 . The method of claim 4 , wherein the overlapping of the plurality of selected sub-signals comprises overlapping the plurality of selected sub-signals after normalizing the plurality of selected sub-signals to a same size.
7 . The method of claim 1 , wherein the selecting of the at least on sub-signal comprises selecting a plurality of selected sub-signals, and wherein the extracting of the target waveform comprises extracting the target waveform by applying weights to the plurality of selected sub-signals, and then overlapping the weighted plurality of selected sub-signals.
8 . The method of claim 7 , wherein the extracting of the target waveform comprises: in response to the target waveform being extracted, determining whether a number of iterations is satisfied; based on the number of iterations not being satisfied, adjusting the weights; extracting a new target waveform by applying the adjusted weights to the plurality of selected sub-signals; and overlapping the plurality of selected sub-signals having the adjusted weights.
9 . The method of claim 8 , wherein the determining whether the number of iterations is satisfied comprises: determining that the number of iterations is satisfied based on a first difference between a representative waveform in a current iteration and a representative waveform in a previous iteration being less than or equal to a first predetermined threshold value; and determining that the number of iterations is satisfied based on a second difference between a quality evaluation result of the plurality of selected sub-signals in the current iteration and a quality evaluation result of the plurality of selected sub-signals in the previous iteration being less than or equal to a second predetermined threshold value.
10 . The method of claim 8 , wherein the extracting of the target waveform further comprises evaluating a quality of the plurality of selected sub-signals, and wherein the adjusting of the weights comprises adjusting the weights based on the quality evaluation result of the plurality of selected sub-signals.
11 . The method of claim 1 , further comprising, based on the target waveform not satisfying the predetermined quality criterion, repeating the measuring of the bio-signal, wherein the selecting of the at least one sub-signal comprises excluding some sub-signals selected during a previous iteration, and including some new sub-signals divided during current iteration.
12 . The method of claim 11 , wherein the selecting of the at least one sub-signal comprises: sorting the plurality of sub-signals in the previous iteration based on at least one of a time sequential order or a descending quality order; and excluding one or more sub-signals from among a bottom of the sorted plurality of sub-signals.
13 . The method of claim 11 , wherein the selecting of the at least one sub-signal comprises excluding a first measured sub-signal and adding a newly acquired sub-signal using First-In First-Out (FIFO).
14 . The method of claim 11 , wherein the selecting of the at least one sub-signal comprises: determining a signal quality of each of the at least one sub-signal; excluding a sub-signal having a lowest signal quality from among the at least one sub-signal; and adding a newly measured sub-signal to the at least one sub-signal.
15 . An apparatus for estimating bio-information, the apparatus comprising: a sensor configured to measure a bio-signal from an object; and a processor configured to: divide the bio-signal into a plurality of sub-signals; select at least one sub-signal for extracting a target waveform from among the plurality of sub-signals; extract the target waveform from the bio-signal by using the at least one selected sub-signal; evaluate a quality of the target waveform; based on the target waveform satisfying a predetermined quality criterion corresponding to the evaluation, determine that the target waveform is a representative waveform of the bio-signal; and estimate the bio-information of the object based on the representative waveform, the bio-information comprising a blood pressure, an arrhythmia, a vascular age, an arterial stiffness, an aortic pressure waveform, and a fatigue level, wherein the processor is configured to: set first weights of first sub-signals from the at least one selected sub-signal having a first quality that meets or exceeds a predetermined quality threshold based on a difference between the first quality and the predetermined quality threshold, and clear second weights of second sub-signals from the at least one selected sub-signal having a second quality that does not meet the predetermined quality threshold.
16 . The apparatus of claim 15 , wherein the at least one selected sub-signal comprises a plurality of selected sub-signals, and wherein the processor is further configured to: set a reference point in each of the plurality of selected sub-signals, and extract the target waveform by overlapping the plurality of selected sub-signals based on the reference point in each of the plurality of selected sub-signals.
17 . The apparatus of claim 16 , wherein the processor is further configured to set, as the reference point, at least one of a minimum point of a waveform of each of the plurality of selected sub-signals, a maximum point of the waveform of each of the plurality of selected sub-signals, a maximum slope point of the waveform of each of the plurality of selected sub-signals, or a tangent intersection point of the waveform of each of the plurality of selected sub-signals.
18 . The apparatus of claim 16 , wherein the processor is further configured to overlap the plurality of selected sub-signals after normalizing the plurality of selected sub-signals to a same size.
19 . The apparatus of claim 15 , wherein the at least one selected sub-signal comprises a plurality of selected sub-signals, and wherein the processor is further configured to extract the target waveform by applying weights to the plurality of selected sub-signals, and overlapping the weighted plurality of selected sub-signals.
20 . The apparatus of claim 16 , wherein upon extracting the target waveform, the processor is further configured to: determine whether a number of iterations is satisfied; in response to the number of iterations not being satisfied, adjust weights applied to the plurality of selected sub-signals; extract a new target waveform by applying the adjusted weights to the plurality of sub-signals; and overlap the plurality of sub-signals having the adjusted weights.

Description

CROSS-REFERENCE TO RELATED APPLICATION(S) This application claims priority from Korean Patent Application No. 10-2021-0145514, filed on Oct. 28, 2021, in the Korean Intellectual Property Office, the entire disclosure of which is incorporated herein by reference for all purposes. BACKGROUND 1. Field The disclosure relates to estimating bio information, and in particular to a method of extracting a representative waveform of a bio-signal, and an apparatus for estimating bio-information based on the extracted representative waveform. 2. Description of the Related Art Due to the aging population, soaring medical costs, and a lack of medical personnel for specialized medical services, research is being actively conducted on IT-medical convergence technologies, in which IT technology and medical technology are combined. Particularly, monitoring of the health condition of the human body is not limited to medical institutions, but is expanding to mobile healthcare fields that may monitor a user's health condition anywhere and anytime in daily life at home or office. Typical examples of bio-signals, indicating the health condition of individuals, include an electrocardiography (ECG) signal, a photoplethysmogram (PPG) signal, an electromyography (EMG) signal, etc., and various bio-signal sensors have been developed to measure these signals in daily life. Particularly, a PPG sensor may estimate blood pressure of a human body by analyzing a shape of pulse waves which reflect cardiovascular status and the like. According to studies on the PPG signal, the whole PPG signal is a summation of a propagation wave propagating from the heart to peripheral parts of a body and reflection waves returning from the peripheral parts of the body. Information for use in estimating blood pressure can be acquired by extracting various features related to propagation waves or reflection waves. However, if the quality of a bio-signal is degraded due to arrhythmia or motion noise in heartbeat, the accuracy in estimating blood pressure may be reduced. SUMMARY According to an aspect of the disclosure, a method of extracting a representative waveform of a bio-signal may include: receiving an input of the bio-signal; dividing the bio-signal into a plurality of sub-signals; selecting at least one sub-signal for extracting a representative waveform from among the divided sub-signals; extracting a representative waveform by using the at least one selected sub-signal; evaluating a quality of the extracted representative waveform; and based on the representative waveform satisfying a predetermined quality criterion corresponding to the evaluation, determining that the representative waveform is a final representative waveform. The dividing of the bio-signal into the plurality of sub-signals may include dividing the bio-signal into the sub-signals based on units of beats. The selecting of the at least one sub-signal may include selecting a predetermined number of sub-signals or sub-signals in a predetermined time interval. The selecting of the at least on sub-signal may include selecting a plurality of sub-signals. The extracting of the representative waveform may include: setting a reference point in each of the plurality of selected sub-signals; and overlapping the plurality of selected sub-signals based on the set reference point in each of the plurality of selected sub-signals. The setting of the reference point may include setting, as the reference point, at least one of a minimum point, a maximum point, a maximum slope point, and a tangent intersection point of the waveform one of the plurality of selected sub-signals. The overlapping of the plurality selected sub-signals may include overlapping the plurality of selected sub-signals after normalizing the plurality of selected sub-signals to a same size. The selecting of the at least on sub-signal may include selecting a plurality of sub-signals. The extracting of the representative waveform comprises extracting the representative waveform by applying weights to the plurality of selected sub-signals, and then overlapping the weighted sub-signals. The extracting of the representative waveform may include: in response to the representative waveform being extracted, determining whether a number of iterations is satisfied; based on the number of iterations not being satisfied, adjusting the weights; extracting a new representative waveform by applying the adjusted weights to the plurality of selected sub-signals; and overlapping the sub-signals having the adjusted weights. The determining whether the number of iterations is satisfied may include: determining that the number of iterations is satisfied based on a similarity between a representative waveform in a current iteration and a representative waveform in a previous iteration being greater than or equal to a first predetermined threshold value; and determining that the number of iterations is satisfied based on a difference between a quality ev