基于呼出气体的口臭检测及疾病筛查的电子鼻研究.pdfVIP

Abstract

characteristiccurve(ROCcurve)todeterminetheconcentrationrangeofVSCswith

differentorganolepticscores.TakingthemeasurementresultsofGC-MSasthe

independentvariablesofthediseasediagnosismodel,adiagnosismodeloforaldiseases

basedonlineardiscriminantanalysisandlogisticregressionmethodwasdesigned,and

thediagnosiseffectsofdifferentmodelswereanalyzedandcompared.

3.TakingthemeasurementresultsofGC-MSasthereferencestandardof

sensorscreening,thesoftwareandhardwaredesignoftheelectronicnosebased

onexhaledgasforhalitosismarkerdetectionanddiseasescreeningwascompleted.

TheconcentrationrangeofVSCsintheexhaledgasofpatientswithhalitosiswas

determinedfromtheresultsofGC-MSmeasurement.TheresultsofGC-MSwereused

asthereferencestandardforthescreeningofelectronicnosesensors.The

electrochemicalsensorswithrange,sensitivity,resolutionandspecificityinaccordance

withtherequirementsofthedetectionwereselectedasthecoredetectionelementsof

electronicnose.Onthisbasis,theairchamberandaircircuitstructureofsensorarray

aredesigned.Theelectronicnosesystembasedonbreathgasforodordetectionand

diseasescreeningisdeveloped.Theinfluenceofhardwareparametersonelectronic

noseperformanceisanalyzedindetail.Thedesignoftheprogramofelectronicnose

slavecomputeranddataacquisitionsoftwareofhostcomputerarecompleted.

4.Thebasicperformanceoftheelectronicnoseistested,thequantitative

algorithmmodelofVSCsbasedonone-dimensionalconvolutionalneuralnetwork

isconstructed,andtheeffectoftheelectronicnoseinthedetectionofactualclinical

samplesisevaluated.

Thedetectionlimit,repeatabilityandlinearityoftheelectronicnoseweretested