matlab毕业设计外文资料翻译

《matlab毕业设计外文资料翻译》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ComplexRidgeletsforImageDenoising1IntroductionWavelettransformshavebeensuccessfullyusedinmanyscientificfieldssuchasimagecompression,imagedenoising,signalprocessing,computergraphics,andpatternrecognition,tonameonlyafew.Donohoandhiscoworkerspioneeredawavelet

2、denoisingschemebyusingsoftthresholdingandhardthresholding.Thisapproachappearstobeagoodchoiceforanumberofapplications.Thisisbecauseawavelettransformcancompacttheenergyoftheimagetoonlyasmallnumberoflargecoefficientsandthemajorityofthewaveletcoeficientsarever

3、ysmallsothattheycanbesettozero.Thethresholdingofthewaveletcoeficientscanbedoneatonlythedetailwaveletdecompositionsubbands.Wekeepafewlowfrequencywaveletsubbandsuntouchedsothattheyarenotthresholded.ItiswellknownthatDonoho'smethodofferstheadvantagesofsmoothne

4、ssandadaptation.However,asCoifmanandDonohopointedout,thisalgorithmexhibitsvisualartifacts:Gibbsphenomenaintheneighbourhoodofdiscontinuities.Therefore,theyproposeinatranslationinvariant(TI)denoisingschemetosuppresssuchartifactsbyaveragingoverthedenoisedsign

5、alsofallcircularshifts.TheexperimentalresultsinconfirmthatsingleTIwaveletdenoisingperformsbetterthanthenon-TIcase.BuiandChenextendedthisTIschemetothemultiwaveletcaseandtheyfoundthatTImultiwaveletdenoisinggavebetterresultsthanTIsinglewaveletdenoising.Caiand

6、SilvermanproposedathresholdingschemebytakingtheneighbourcoeficientsintoaccountTheirexperimentalresultsshowedapparentadvantagesoverthetraditionalterm-by-termwaveletdenoising.ChenandBuiextendedthisneighbouringwaveletthresholdingideatothemultiwaveletcase.They

7、claimedthatneighbourmultiwaveletdenoisingoutperformsneighboursinglewaveletdenoisingforsomestandardtestsignalsandreal-lifeimages.Chenetal.proposedanimagedenoisingschemebyconsideringasquareneighbourhoodinthewaveletdomain.Chenetal.alsotriedtocustomizethewavel

8、et_lterandthethresholdforimagedenoising.Experimentalresultsshowthatthesetwomethodsproducebetterdenoisingresults.Theridgelettransformwasdevelopedoverseveralyearstobreakthelimitationsofthewavelettransform.The2D