資源描述:
《基于非采樣contourlet變換和圖論的醫(yī)學(xué)圖像侵害》由會員上傳分享,免費在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、山東大學(xué)碩士學(xué)位論文ABSTRACTImagesegmentationistheimportantpartinimageprocessingtechnology.Withthedevelopmentoftheinformationtechnology,Medicalimagesegmentationplaysamoreandmoreimportantroleinmedicalapplications.Medicalimagealgorithmefficiencyhasimportanttheoreticalresearchmeaningandcommercialva
2、lue.Inmedicalimages,thereoftenwillbeunnecessarynoisedistributioninfocalregion;thiswillincreasethedifficultyofthedoctortoaccuratelydiagnosisthedisease.Whenthemedicalimageusingageneralsegmentationmethodtosegmentmedicalimage,usuallyneedtodenoisetheimagefirstly,butthedenoiseprogresswillmak
3、ethesegmentationprocesstimelonger,thisrestricttheapplicationofmedicalimagesegmentationtechnology.Fortheabovephenomenon,thispaperproposesthemedicalimagesegmentationbasedonnon—subsampledContourlettransform(NSCT)andgraphtheory.Firstly,usenon—subsampledContourlettransformtheimage,gettheCon
4、tourletcoefficientofmultiscaleanddirectionalinformation.Non—subsampledContourlettransformisdecomposedfastalgorithmofimagewiththedirectionandwiththeshiftinvariance.Imagesegmentationbasedongraphtheoryisahottopicthisyear,becausethegraphtheoryisamaturetheoryinresearchandapplicationrelative
5、ly,thispapermainlyintroducestwokindsofsegmentationmethodbasedonGraphTheory:normalizedcutsegmentationandisoperimetricsegmentationmethod,becausethecuttingmethodbasedonGraphTheoryisgenerallyalargeamountofcalculation,andaffectedbynoise,SObeforethesegmentation,weusingNon-subsampledContourle
6、ttransformextractthelow-frequencyinformationoftheimage,researchshowsthat,thelowfrequencycoefficientsretainthemostoftheinformationintheimage,theunnecessarynoiseisfiltering,anditreducestheredundantinformationinimage.Sointheflowingsegmentationprocess,wecansavealotoftime,andgetagoodsegment
7、ationresults.Theproposedalgorithmisverifiedinthespecialtyofmedicalimage.The山東大學(xué)碩士學(xué)位論文experimentalresultsshowthat,theproposedalgorithmobtainedgoodresultsandgreatlyshortentheprocessingtime,whichhasanimportantsignificanceinpracticalapplication.Theexperimentalresultsshowthat,therecogniti