Example: bankruptcy

Signal Processing Overview of Ultrasound Systems for ...

WhitePaperSPRAB12 November2008 SignalProcessingOverviewofUltrasoundSyst emsforMedicalImagingMurtazaAli, :thefrontend(thatincludestransmission,re ception,anddigitizationofultrasoundsigna l),followedbybeamforming(tofocusonaparti cularscanline),themidend(thatperformsdem odulation,envelopedetection,andcompressi on),andfinally,thebackend(thatperformsva riousimageenhancementandrenderingfunctio ns).TheprocessingrequiredinDopplermodeis describedand,finally,theprocessingandiss uesrelatedto3 ,obstetrics,gynecology,abdominalimaging, , ,( ,inguidinginterventionalprocedures). ,digitalsignalprocessors(DSP) (SOC) [20]. , (DAS) , ,whichiscomposedofimageenhancement,noise reduction, , ,Section8brieflyintroducesthebasicconcep tsusedfor3 November2008 SubmitDocumentationFeedback2 UltrasoundSystem:BasicConceptsRegion of InterestScan LinesFocal ,bynomeans, , ,ultrasoundsystemsoperateinthe2 MHzto20 MHzfrequencyrange,althoughsomesystemsare approaching40 MHzforharmonicimaging[3].

1 Introduction Introduction www.ti.com Ultrasound is one of the most widely used modalities in medical imaging. Ultrasound imaging is regularly used in cardiology, obstetrics, gynecology, abdominal imaging, etc. Its popularity arises from the fact that it provides high-resolution images without the use of ionizing radiation. It is also mostly ...

Tags:

  Introduction, Introduction introduction www

Information

Domain:

Source:

Link to this page:

Please notify us if you found a problem with this document:

Other abuse

Transcription of Signal Processing Overview of Ultrasound Systems for ...

1 WhitePaperSPRAB12 November2008 SignalProcessingOverviewofUltrasoundSyst emsforMedicalImagingMurtazaAli, :thefrontend(thatincludestransmission,re ception,anddigitizationofultrasoundsigna l),followedbybeamforming(tofocusonaparti cularscanline),themidend(thatperformsdem odulation,envelopedetection,andcompressi on),andfinally,thebackend(thatperformsva riousimageenhancementandrenderingfunctio ns).TheprocessingrequiredinDopplermodeis describedand,finally,theprocessingandiss uesrelatedto3 ,obstetrics,gynecology,abdominalimaging, , ,( ,inguidinginterventionalprocedures). ,digitalsignalprocessors(DSP) (SOC) [20]. , (DAS) , ,whichiscomposedofimageenhancement,noise reduction, , ,Section8brieflyintroducesthebasicconcep tsusedfor3 November2008 SubmitDocumentationFeedback2 UltrasoundSystem:BasicConceptsRegion of InterestScan LinesFocal ,bynomeans, , ,ultrasoundsystemsoperateinthe2 MHzto20 MHzfrequencyrange,althoughsomesystemsare approaching40 MHzforharmonicimaging[3].

2 Inprinciple, , , , ,asetoftransducerelementsareenergizedwit hasetoftime-delayedpulsestoproduceasetof soundwavesthatpropagatethroughtheregiono finterest, , , , ,inwhichtheultrasoundsystemusuallyoperat es,doesnotmatchthedisplaycoordinatesyste ms,acoordinatetransformation,calledscanc onversion, November2008 SignalProcessingOverviewofUltrasoundSyst emsforMedicalImaging3 SubmitDocumentationFeedbackHVTransmitter HVMUXS ignalChainREFT ransmitBeamformerFront EndBackEndReceiveBeamformerBeamformerCon trol UnitPreprocessingSpectralDopplerProcessi ngB ModeProcessingColor/PowerDopplerProcessi ngScan ConversionPost ProcessingCore & LineAC/DC SupplyWith Green ModeControllerMain Power SupplyBacklightTouchscreenSystemPowerSup plyVoltT/RSwitchesMemoryInterfacesMMC/SD IOIDE/ATAU ltrasound OS/UIMidClockingTransducerTx BufferAmpPassiveLPFCW (analog)BeamformerLow PassFilterDisplayAudio OutputAudioAmpLEGENDP rocessorInterfaceAmplifierLogicPowerADC/ DACO therAmplifier + FilterPGALNAA mplifier StageVCAADCA nalog , 100 Vand (Tx/Rx)switch, , , (VCA)

3 Beforebeingsampledbyanalog-to-digitalcon verters(ADC).TheVCAisconfiguredsothatthe gainprofilebeingappliedtothereceivedsign alisafunctionofthesampletimesincethesign alstrengthdecreaseswithtime( ,ithastraveledthroughmoretissue). , (ASIC),field-programmablegatearray(FPGA) , ,whichdeterminestheinput/output(I/O) :BasicConceptsOncethedataisbeamformed,de pendingontheimagingmodes, , (Brightness)mode, , ,finally, ,anwindowedfastFouriertransform(FFT) ,ASIC, , ,modernSOCsthatincludeDSP,RISC, ,thissectionhasconcentratedondigitalbeam forming,whichiscommonforpulsewave(PW) ,inthesesystems, , (CW) , ,thedynamicrangerequirementbecomesveryhi ghtouseADCtodigitizethereflectedultrasou ndsignalandmaintainenoughsignaltonoise(S NR) , (usuallyinKHzrange) sultrasoundsystem. A-mode(Amplitude) , B-mode(Brightness)imagingisthesameasA-mo de, ,multiplesetsofpulsesaregeneratedtoprodu cesoundwavesforeachscanline,eachsetofpul sesareintendedforauniquefocalpointalongt hescanline.

4 ForCW(ContinuousWave)Doppler, , ,thebloodflowinveinscanbeestimatedusingt heDopplerfrequency(seeSection7fordetails ).However,sincethesensoriscontinuouslyre ceivingdatafromvariousdepths,thevelocity locationcannotbedetermined. ForPW(PulseWave)Doppler, ,thevelocitylocationcanalsobedetermined. ForColorDoppler, InPowerDoppler,insteadofestimatingtheact ualvelocityofthemotion, (upto1000pulsespersecond),thisisusefulin assessingratesandmotionandisstillusedext ensivelyincardiacandfetalcardiacimaging. HarmonicImagingisanewmodalitywheretheB-m odeimagingisperformedonthesecond(orpossi blyother)harmonicsoftheimaging[3].Duetot heusualhighfrequencyoftheharmonic, ,duetohigherloss, Elasticity/StrainImagingisanewmodalitywh eresomemeasuresofelasticity(likeYoung smodulus)ofthetissue(usuallyundercompres sion)isestimatedanddisplayedasanimage[17 ],[18].

5 , :specular,diffusiveanddiffractive[24]. ,thereflectionprocesscanbeapproximatedas anincidentraywiththescatteredwavefrontfo llowingtheshapeoftheobject..Bycentrallim ittheorem, , (specularscattering), ,thoughmodeledwithadistributionovertheim age, , (likeboundarydetection) , ( )()MTF f , zexp ( f )zg=(1)( f )( f )i ( f )gab= --(2)y(f )f0 1aaa=+(3)( )( )( )fkff0 Ebb=+(4)( )( )( )P f , zP f MTF f , z0=(5) (MTF)writtenasafunctionoffrequency,f,and depthofpenetration,z,inthefollowingmanne r[24]:wherek0=2 f/c0,c0beingthevelocityofsoundusuallytak enatthecenterfrequencyofthespectrum, (f)isanexcessdispersiontermresultinginfr equencydependentphasevelocity,and (f)isthefrequencydependentattenuationfol lowingpowerlaw( 0, 1,andyareconstants).IfapressureP0(f)isap pliedatdepthz=0,thepressureatanydepthzis thengivenby:TheinverseFouriertransformof MTF(f,z) ,thefollowingcharacteristicscanbederived : ,theattenuationislow( )inblood.

6 Velocitydispersion,thoughsmall,mayneedto beconsideredforwidebandsignalsasisthecas eofmodernultrasoundtransducers. Intimedomain, ,responseisimpulselikeandwithdepth,itspr eadsovertime. Infrequencydomain,thereisaslightdownward shiftofpeakintheresultingfrequencyspectr umwithdepthofpenetrationwhenashaped( ,Gaussian) , :linear,curved, , , , , ,thesignalflexisusuallysolderedtoagroupo fwiresinacablebundlethatisconnectedtothe systemviaastandardizedconnector( ,ZIF156). , , , , , , , ,alternatetransducerstructuresusingcapac itivemicro-machinedultrasoundtransducers (CMUTs)isanactiveareaofresearch[33].8 SignalProcessingOverviewofUltrasoundSyst emsforMedicalImagingSPRAB12 PointXZidq( )222sinRxx Rfpii fptciq+-=(6)( )22Rx2 xRsinmaxmax fpfptmaxcq++=(7)( )()ptp tiit=-(8) ,theprocessofsteeringandfocusingthesound beaminanultrasoundsystemiscommonlyreferr edtoasphasedarraybeamforming[26].

7 Inthismodeofoperation, , ,p(t).Thepropagationtime,ti,fortheithpie zo-electricelementtothefocalpointcanbewr ittenas:wherexiisthex-axiscoordinateofth eithpiezo-electricelement,Rfp,istheradia ldistancefromtheorigintothefocalpoint, ,istheangleofthefocalpointwithrespecttot hez-axis, , , ,themaximumpropagationtimeforagiven(Rfp, )configuration( ,focalpoint) ,thepulsewaveformfortheithpiezo-electric elementbecomes:where I=tmax-tiSPRAB12 November2008 SignalProcessingOverviewofUltrasoundSyst emsforMedicalImaging9 SubmitDocumentationFeedback( )()NNr tAA p tritjri tji 1 j 1tt=--== (9)( )( )2cos2 Rsinxxfpii2 Rfpticqq-+ (10)( )sin,xciq-( )2cos2, , , ,r(t),fromagivenfocalpointafterithasbeen timelignedcanbewrittenas:whereAriistheap odizationappliedtoeachreceivesignal,Atji stheapodizationappliedtoeachtransmitsign al, riisthetime-delayvalueappliedtoeachrecei vedsignal,and , , ,Equation6cannotbeseparatedintotermsthat justdependonRfpand.

8 [28].Thelinearterm,isoftenreferredtoasth ebeamsteeringcomponentbecauseitonlydepen dsupon , x,whereiisthechannelnumberand , , ,detection,andlogcompression(seeFigure6) . (TFC) (conventionalimaging)orthesecondharmonic (harmonicimaging). , (conventionalversusharmonic),orchangesin thecenterfrequencieswithtime( ,aphysicalphenomenoninthetissuethatresul tsinachangeoffrequencywiththedepthofpene tration).However,thisoperationismorecomp lexcomparedtothealternativebelow. , [24].Theactualdynamicrangeofthereceiveds ignaldependsontheADCbits,theTGCamplifier usedinthefrontend, (usually7or8bits). [16].Thecodedexcitationisusedtoincreaset hedepthofpenetrationandincreasesensitivi tywithoutincreasingthetransmitpulsepower ,orequivalently, Codedexcitation:Severalcodedsequenceslik eBarkerandGolaysequences[10]havebeenused .

9 Chirps:Thisisborrowedfromradarwherethetr ansmitwaveformconsistsofalinearsweepoffr equencymodulatedpulseofagivenduration[11 ].Asisclearfromthechoiceofcodedsequences , ,thenthereceivedsignalneedtobefilteredth roughthecorrespondingmatchedfilter(orseq uence).Toreducetheeffectofchangingcenter frequencywiththedepthofpenetration, ()()1y(i )x(i ) (i )y(i 1) ,1(i ) f x(i ) y(i 1)aaa=+-+=--(11) ,theyareweightedandadded[24]. , ,andwhensummed, (TGC), , , ( ,gallbladder).Insuchcases,itiscommontopr ovideadditionalgaincompensationbothinthe axialandlateraldirections[15],[25].These gaincompensationscanbeunderusercontrol,b utitisalsopossibletoautomaticallydetects uchanechoicregionsandprovideanautomaticg aincontrol(AGC) , , (FIR/IIR) ,asingletapmotionadaptiveIIRfilteringcan beimplementedusingEquation11[22]:Here,y( I)istheoutputpixelvalueattimeI,x(I)thein putpixelvalue, Data in Polar CoordinatesZoomed VersionData Converted to Cartesian CoordinatesExtrapolated Cartesian Point (Outside Region)Interpolated Cartesian PointOriginal Polar (Spatial) , (forlinearprobes)orinpolarcoordinates(fo rcurvilinearorphasedarrayprobes).

10 , (2X2interpolation).Somehigh-endsystemsma yuseinterpolationbasedon16nearestneighbo rs(4X4interpolation).Thisisamulti-imagem ethodusedforspecklereduction[24].Theidea isthatbycombiningviewsofthesameobjectfro mdifferentangles,eachhavinganuncorrelate dspecklerealization, ,compensatingthemfortranslation/rotation (involvesmultiplyingwitha2 Dmatrix) November2008 SignalProcessingOverviewofUltrasoundSyst emsforMedicalImaging13 SubmitDocumentationFeedback0 degree acquisition20 degree acquisition-20 degree acquisitionWeighted averagingCompounded ( x, y )w( s,t )f ( x s, y t )sa tb=++= -= - (12) , (illustratedas+20 ,0 and-20 ). (duetolessreflection);therefore, (-1,-1)w(-1,0)w(-1,1)w(0,-1)w(0,0)w(0,1) w(1,-1)w(1,0)w(1,1) November2008 SubmitDocumentationFeedback221x1yh( x,y ) g ( x,y ) (exp()) (exp()),2D222222g ( x) g (y )1D1Dp sp sss==- -= (13) f22f ( ) ( ), wherexyx yfff( x,y ) f( x 1,y ) andf( x,y ) f( x,y 1)xy =+@+ =--=-- (14) [1].


Related search queries