简介:Thequadratictransformationmethodisproposedtoestimatethetrabecularspacing(Tb.Sp),animportantindexforosteoporosisdiagnosis.Theperformanceofthisalgorithmwasinvestigatedbyscattermodel,two-dimensionfinitedifferencetimedomain(2D-FDTD)simulationandinvitroexperimentsofbovinecancellousbonespecimens.Comparingwiththeotherfourmethods—autoregressivecepstrum(AR),adaptivefilter-autoregressivecepstral(AFAR),inversefilter-autoregressivecepstrum(InvAR),andsimplifiedinversefiltertracking(SIFT),quadratictransformationismuchmorestableandaccurate.TheresultsdemonstratedthatquadratictransformationisagreatalgorithmforTb.Spestimation.
简介:Thecontributionsofthestaticanddynamicarticulatoryinformationtospeechrecognitionwereevaluated,andtherecognitionapproachesbycombiningthearticulatoryinformationwithacousticfeatureswerediscussed.ArticulatorymovementswereobservedbytheElectromagneticArticulographicSystemforreadingspeech,andthespeechsignalswererecordedsimultaneously.First,weconductedseveralspeechrecognitionexperimentsbyusingarticulatoryfeaturesalone,consistingofanumberofspecificarticulatorychannels,toevaluatethecontributionofeachobservationpointonarticulators.Then,thedisplacementinformationofarticulatorydatawerecombinedwithacousticfeaturesdirectlyandadoptedinspeechrecognition.Theresultsshowthatarticulatoryinformationprovideswithadditionalinformationforspeechrecognitionwhichisnotencodedinacousticfeatures.Furthermore,thecontributionofthedynamicinformationofthearticulatorydatawasevaluatedbycombiningtheminspeechrecognition.Itisfoundthatthesecondderivativeofarticulatoryinformationprovidedquitelargercontributiontospeechrecognitioncomparingwiththesecondderivativeofacousticalinformation.Atlast,thecombinationmethodsofarticulatoryfeaturesandacousticoneswereinvestigatedforspeechrecognition.ThebasicapproachisthattheBayesianNetwork(BN)isaddedtoeachstateofHMM,wherethearticulatoryinformationisrepresentedbytheBNasafactorofobservedsignalsduringtrainingthemodelandismarginalizedasahiddenvariableinrecognitionstage.ResultsbasedonthisHMM/BNframeworkshowabetterperformancethanthetraditionalmethod.
简介:一个修改Parzen窗户方法,在低频率使分辨率高并且把光滑放在高频率,被建议获得统计模型。然后,当长句子被处理时,利用统计模型的一个性分类方法被建议,它有性分类的98%精确性。由男声音和女性表示的分离,与不同情感训练样品的讲话的平均数和标准差被用来创造相应情感模型。然后在测试样品和沥青的统计模型之间的Bhattacharyya距离,在沥青的speech.The正规化为情感识别被利用因为男声音和女声音也被考虑,以便说明他们直到一个一致空格。最后,讲话情感识别实验基于K最近的邻居显示出那,81%的正确的率被完成,在它仅仅是73.85%if的地方,传统的参数被利用。
简介:Usingtheextremedifferenceofself-similarityandkurtosisatlargelevelscaleofwavelettransformapproximationbetweenthePTFM(PulseTrainsofFrequencyModulated)signalsanditsreverberation,afeature-basedmatchedfiltermethodusingtheclassify-before-detectparagriamisproposedtoimprovethedetectionperformanceinreverberationandmultipathenvironments.Processingthedataoflake-trailsshowedthattheprocessinggainoftheproposedmethodisbiggerthanthatofmatchedfilterabout10dB.Inmultipathenvironments,detectionperformanceofmatchedfilterbecomebadlypoorer,whilethatoftheproposedmethodisimprovedbetter.Itshowsthatthemethodismuchmorerobustwiththeeffectofmultipath.
简介:Aspeechsignalprocessingandfeaturesextractingmethodbasedoncomputationalauditorymodelisproposed.Thecomputationalmodelisbasedonpsychological,physiologicalknowledgeanddigitalsignalprocessingmethods.Ineachstageofahearingperceptionsystem,thereisacorrespondingcomputationalmodeltosimulateitsfunction.Basedonthismodel,speechfeaturesareextracted.Ineachstage,thefeaturesindifferentkindsoflevelareextracted.Afurtherprocessingforprimaryauditoryspectrumbasedonlateralinhibitionisproposedtoextractmuchmorerobustspeechfeatures.Allthesefeaturescanberegardedastheinternalrepresentationsofspeechstimulationinhearingsystem.Therobustspeechrecognitionexperimentsareconductedtotesttherobustnessofthefeatures.Resultsshowthattherepresentationsbasedontheproposedcomputationalauditorymodelarerobustrepresentationsforspeechsignals.
简介:Multitaper光谱比传统的periodogram有更低的变化。到吵闹的信号光谱比率(NNSR)的噪音光谱和噪音从吵闹的信号的multitaper光谱被估计;为计算掩盖阀值的噪音的预先提高的讲话被获得由光谱振幅减法方法,其获得是NNSR的功能;最后的提高的讲话被与合并掩盖阀值的噪音的psychoacousticalweighting规则压制吵闹的讲话的Fourier光谱获得。因为multitaper光谱的低变化特征,一个修改偏移量公式被建议计算掩盖阀值的噪音,因此,有这修正的重建的讲话在MBSD有改进(修改吠叫光谱失真).When不到一统治到psychoacousticalweighting的最大的限制进一步被建议,更高输入SNR(0dB)是,部分SNR和全面SNR有的更多的改进。非正式的听测试证明几乎为由建议方法的提高的语音处理没有很少讲话失真,背景噪音很并且没有音乐的噪音被减少。
简介:InordertoreconstructcomplicatedtemperaturefieldsmoreaccuratelybyacousticCT,areconstructionalgorithmbasedonMarkovradialbasisfunctionandTikhonovregularizationisproposedandnamedasMTRalgorithm.Withthealgorithm,theacousticvelocityfieldinamediumisapproximatedbyalinercombinationofMarkovradialbasisfunctions,theacoustictravel-timesovermulti-pathsandtheTikhonovregulationareusedtoreconstructtheacousticvelocitydistribution,andthenthetemperaturedistributioniscalculatedbyusingtherelationshipbetweenacousticvelocityandtemperature.Thetemperaturefieldmodelswithonehotspot,threehotspotsandfivehotspotsarereconstructedbyusingsimulationdata.ReconstructionresultsshowthattheMTRalgorithmcanreconstructthehottemperature,especiallythehotpositionaccurately.AnexperimentsystemfortemperaturedistributionmeasurementbyacousticCTisdeveloped.ThecapabilityofacousticCTtodetectahotspotcreatedbyelectricheatersinanexperimentalsilofilledwith1200kgsoybeansistestedbyusingtheMTRalgorithm.Inthereconstructiontemperaturefield,thehotpositioncanbedeterminedcorrectlyandthetemperatureerrorofthehotspotis1.3%.ItisthusclearthattheMTRalgorithmhasagoodcapacityofreconstructingcomplextemperaturefields,andcanbeexpectedtobeusedintemperaturemonitoringforactualstoredgains.
简介:拿估计到达的方向被介绍的进报道的在数组元素之中的相互的联合的一个multi-subarraysubspace试穿方法。一致线性数组的相互的联合矩阵与banded被建模对称的Toeplitz矩阵。根据相互的联合矩阵的DOF(自由的度),在二个方面的数组元素的部分被忽视。精神算法的理论特征,适合算法的subspace之一,被学习。为大N和一致线性数组,MCLS精神评价错误是asymptotically,联合,与零个工具散布的Gaussian,和它的协变性表达式被获得。它从模拟被知道在那里存在当当它的变化有一些增加时,相互的联合不存在时,有为数组元素和信号来源的某些数字的最低评价错误的subarray,和它估计到达的方向的性能接近理想的状况。方法与高分辨率被用于深测术的sidescan声纳,并且好结果被获得了。以增加数组元素的数字的成本,方法能减少在数组元素之中的相互的联合的爱。
简介:Afocusingtransformbasedminimumvariancemethod(FTMV)isproposedforestimatingtheazimuthanglesofwidebandsignals.TheestimationerrorofDirectionofArrival(DOA)inthetraditionalMinimumVariance(MV)methodduringthearraydirectionchangingwiththeshipcanbeeliminatedbyfocusingthesamplecovariancematricesinwidebandontoasinglenarrowband,andaccumulatingthebeamformeroutputsonazimuth.Then,theazimuthanglesofsignalscanbeestimatedaccurately.ExperimentalresultsshowthattheFTMVmethodismoreefficientandrobustthantheMVmethodinthiscase.
简介:以便减少磨擦力量并且在低速度消除一个技工系统的粘住滑倒现象,一个方法基于超声的微开车技术被建议改变摩擦特征。顺时针方向激动并且在超声的致动器的二个纵的夹钳门栓的震荡器上驾驶点的逆时针方向的显微镜的椭圆的运动将产生超声的润滑油行动;而且,磨擦能被调整震荡器的颤动的振幅活跃地控制。为磨擦控制的试验性的安装用航空的指南,力量传感器和一低速度时刻被设计motor.Fuzzy控制理论被使用进这个系统。实验显示磨擦力量大部分被减少了,试验性的系统的运动是稳定的。当超声的致动器和负担is3.8kg和马达的全部的团正在开车时,磨擦系数仅仅是大约0.0053速度是0.5mm/s。
简介:Cyclostationary健全的地是nonstationary声音地,压力信号严重在被调制的一种特殊类型,边带在它的光谱存在。重建的健全的地不能在常规Nearfield声学的雷射摄影术考虑cyclostationary特征(不)过程。根据平面cyclostationary不,cyclostationary不基于边界元素方法被建议它能被利用与复杂表面分析散热器。用秒顺序代替Fourier的变换周期的统计,周期光谱密度(CSD)功能被用作重建的物理数量在不建议了技术,而不是光谱或力量光谱压力的密度信号。由CSD功能的解调能力的优点,重建的CSD能有效地分别地表示modulating和搬运人波浪的信息。模拟和实验说明这cyclostationary的有效性和精确性不技术满足工程的请求。
简介:在医药Doppler超声系统,是的一个高通行证的过滤器通常采用了过滤墙喧嚷部件,将移开未经触动的Doppler超声血表明的低速度血flow.To摘录的信息,一条新奇途径基于空间地选择的噪音过滤被建议。墙信号被空间地选择的噪音过滤第一从小浪估计空间关联性质。然后,墙喧嚷被降噪消除剩余血流动信号的技术的小浪阀值确切获得。最后,未经触动的血流动信号被从混合信号减去墙信号完成。这条途径被用于模仿的计算机并且在vivo颈动脉Doppler超声信号。实验结果证明基于的小浪空间来临能确切提取血流动信号,并且比高通行证的过滤的在吝啬的绝对错误完成大约45%更低的结果。这条途径被期望是一个有效方法在Doppler超声系统移开墙喧嚷。
简介:InordertoimprovetheperformanceofdeceptiondetectionbasedonChinesespeechsignals,amethodofsparsedecompositiononspectralfeatureisproposed.First,thewaveletpackettransformisappliedtodividethespeechsignalintomultiplesub-bands.Bandcepstralfeaturesofwaveletpacketsareobtainedbyoperatingthediscretecosinetransformonlogarithmicenergyofeachsub-band.ThecepstralfeatureisgeneratedbycombingMelFrequencyCepstralCoefficientandWaveletPacketBandCepstralCoefficient.Second,K-singularvaluedecompositionalgorithmisemployedtoachievethetrainingofanover-completemixturedictionarybasedonboththetruthanddeceptivefeaturesets,andanorthogonalmatchingpursuitalgorithmisusedforsparsecodingaccordingtothemixturedictionarytogetsparsefeature.Finally,recognitionexperimentsareperformedwithvariousclassifiedmodules.Experimentalresultsshowthatthesparsedecompositionmethodhasbetterperformancecompariedwithconventionaldimensionreducedmethods.Therecognitionaccuracyofthemethodproposedinthispaperis78.34%,whichishigherthanmethodsusingotherfeatures,improvingtherecognitionabilityofdeceptiondetectionsystemsignificantly.
简介:Acoherence-basedcorrectionmethodwasproposedinordertoimprovethelateralresolutionandenhancethecontrastofmedicalultrasoundimaginginthepresenceofphaseaberration.Theaveragedcoherencefactorwasproposedatfirstandusedasametrictoevaluatephaseaberrationcorrection.Bymaximizingtheaveragedcoherencefactor,thetimedelayparameterofeachchannelwasadjusted.AnewsetofcoherencefactorswascalculatedandthecorrecteddatawasoptimizedtoformthefinalB-modeimage.Thesimulationsonpointtargetsandacystphantomshowedthattheproposedmethodoutperformedthenearestneighboringcrosscorrelationmethodandconventionalcoherence-weightingmethod,andthelateralresolutionandcontrastratiowasimprovedbyapproximately0.24mmand18dBrespectively.Theproposedmethodcombinedtheadvantagesofphaseerrorcorrectionandcoherence-weighting,whichcouldimproveimagingqualitieseffectivelyinmedicalultrasound.