Investigation on the effect of Womersley numberECG and PPG features for cuff less blood pressure estimation using machinelearni.docx

《Investigation on the effect of Womersley numberECG and PPG features for cuff less blood pressure estimation using machinelearni.docx》由会员分享，可在线阅读，更多相关《Investigation on the effect of Womersley numberECG and PPG features for cuff less blood pressure estimation using machinelearni.docx（11页珍藏版）》请在第一文库网上搜索。

1、FISFVIERContents1istsavai1ab1eatScienceDirectBiomedica1Signa1ProcessingandContro1journa1homepage:cInvestigationontheeffectofwomers1eynumber9ECGandPPGfeaturesforcuff1essb1oodChockforupdatespressureestimationusingmachine1earningGeerthyThambiraj,1JmaGandhi,UmapathyManga1anathan,VJeyaMariaJose,M.AnandDe

2、pr(mentofIns(runena(ionandContno1Engineering1Naiona1IMitu(eoechno1ogy.Tirchirapa11iJndiaartic1einfoArtic1ehistory:Re(XiVCd230CIobeI2019Reccivcdinrcviscdfonn13Fcbruarj2023ACCCPtCd28Mairh2023Keywords:ContinuouscufncssMmdpressurcRandomforcstWomcis1cyparamctcrECGfcaturesGenetica1gorithmabstractObjective

3、:Regu1ationandinhibitionofhighbkxx1pressure.knownashypertensionareintrica(e.andiidemandsacontinuous.accuraieb1x1pressurcmcasuremcntsystcm.A1hhccxistingcontinuousnoninvasivctcchniqucsowncha11cngcssuchascxactp1accmcntofthcsensor,reconstructionofarteria1pressurefromfingercufffrequentandsubjectbasedca1i

4、bration.Thispaperpresentsana1gorithmbasedonncwtimc-iiomaintcaturesfbrcontmuousb1oodpressurcmonitoringwhichisccia1inin(ensivccarcUnitsandcanbcuscdtoPNdicicardiovascu1arai1ments.Methods:Here.WeproposethemethodtoesiimateBpthaiextractsinfbrmativeteatures1ikeWmers1eynumber().QRS.QTcinter,a.SDirromEcGandp

5、pGsigna1sandregressiontechniqueswhichareemp1oyedtoesiinateb1oodressurecontinuous1y.Performancemetrics1ikeMAE.RMSE.nbias&95%CIareconsideredtova1idatetheProPOmethod.Toexp1orethere1evanceofproposedphysio1ogica1fcaturcswiththcb1(1pressurc,gcnctica1gorithm(GA)withthcrandomforcstmodc1iscmp1oycd.Resu1ts:Si

6、gnificantfeatures1ikca1pha4QRScomp1cx,QTintcrra1SD1hcartratcarcacquired.Thebestoptima1featuresetfromGArcduccdthcMAEfrom13.20to9.54mmHg,9.91to5.48mmHg,7.71to3.37mmHgibrSBP.DBPandMBPrcspcc(ive1y.Conc1usion:Mode1bui1dwithECGandPPGtime-domainfeaturesoutperformthemode1trainedwithPPGfeaturesa1one.Resu1tsO

7、txaincdfromGAvaIidatcsIhcsignificanccofEcGfcaturcsconrcIationwithBRSignicancc:Identifying(hcass,methodisa1soawe11-estab1ishedmethodtoestimateBPUsingamaximumamp1itudea1gorithmJnthistechniqucahesma11osci11ationsareextracted,anditsenve1opesarecomputedwhichgivetheestimationofsysto1icb1oodpressure(SBP)an

8、ddiasicb1oodpressure(DBP)2.However,boththetechniquesprovidetheintermittentBP.constricttheb1otif1owandrequiretrainedtechnicians.Inaddition,thevo1umec1ampingmethodandthetonometricmethodusesthefingercufftoreconstcttheb1oodpressureva1ues.Howeverjhep1acemeiitofthesensorontheexact1ocationofthearteryischa1

9、1engingintonometr),P1-Thetechniquesmentionedabovehavetheirdrawbacks,whichresu1tintheemergingprincip!eca11edpu1sewaveve1ocity(PWV)basedBPmeasurement.Forins(ance,thecufT-basedmeasurementOfb1oodpressurecausesdiscomfbrtinthecaseOfambu1atorymonitoring,inappropriatecuffsizingandprovidesanintermittentmeasu

10、rement.Hence,cuff1essBPassessmentbasedonPu1setransittime(PTT)isevo1vedtoOvercomethedemeritsassociatedwiththecuffbasedmeasurement.Thethreemostciteda1gorithmsbasedonPTTpavedthewayforcontinuousandcuff1essb1oodpressuremeasurement.Cheneta1.estimatedtheSBPbyobtainingchangeinpu1setransittimea1ongwithinterm

11、ittentca1ibrationva1ues.However,heobservedareductioninaccuracyWhentheca1ibrationinterva1is1onger4.Pooneta1repIacedtheC1aSIiCmOdU1USWiIhHUgheSCqUaIionandCa1CUIaIedSBPandDBP5.AIthOUgh1hCabovetwoa1gorithmsestimatedb1oodpressure,thediameterissetasideasaconstantparameter.In2015,Dinge1a1.proposedthenewind

12、icator,photop1ethysmogramintensityratio(PIR)whichref1ectstheVasomotortonexhangeinarteria1diamcterasWe11asitcou1dIrackbothhighand1owfrequencyvariationsinBP6.Inphysio1ogica11erms.b1oodpressureisregu1atedbycardiacoutputandtota1periphera1resistance,whichinc1udevesse1diameter,b1oodviscosity,andvesse11engthandb1oodvo1ume.Therefore,anove1Bpa1gorithmbasedonMorganaiK1Ki1eyexpressionasaprincipaIequationwasmode11edwhichre1