基于小波變換的頭肩序列圖像編碼技術(shù)研究.pdf

1、大慶石油學院碩士學位論文基于小波變換的頭肩序列圖像編碼技術(shù)研究姓名：魏玉芬申請學位級別：碩士專業(yè)：計算機應(yīng)用技術(shù)指導教師：唐國維20060313ResearchonWaveletTransform—basedheadshoulderSequenceCodingtechniquesABSTRACTWiththeageofinformationcoming，theresearchofimagecodingmethodhasbecomeone

2、ofthemostactivefieldsininformationtechnulogybywayofcomputercommunication，informationstore，InternetcommunicationtechnologyetcEspeciallyin21stcenturywiththedevelopmentofelectronicandcommunicatetechnologyitiSpossibletoreali

3、zetheVideotelephone，meetingⅣ，signalTv，informationhighwayetcOnthisoccasion，itinevitablybecomesoneofthemaintaskstoseekaftertheeffectiveimagecodingmethod，anditwillbecomeonehotpointofresearchathomeandabroad，Thetanalysisofwav

4、eletiSdevelopinginthe1980’StodealwithunbalancedsignalsOnthebasisofmuhiresolutionitovercomet11edefaultofFourieranalysisT11estudyofwaveletbasedimageandvideocompressionhasdrawnmuchattentioninthisfieldTheimageandvideocodingb

5、asedonwavelettransformiSstudiedinthisthesismadeimprovementstoitsclassicalgorithmsandCOnductedexperimentscertification。FirstlyIntroducingthetheoryofwavelettransfoITII，theconceptofmulti—resolutionanalysis，andtheMallatalgor

6、ithmThesignalofNossiniscomposedanddecomposedbyM棚latalgorithmTheadvanceofwayelettransformiSfurtheranalyzedWjvelettransformCanch赫2ethedistributionofimageenergyandmaketheenergyconcentratingSoitCanavoidthediamondeffectanduti

7、lizethefeatureofprogressivetransmissionEZWalgorithmsforthefollowingaspectsoftheproposedimprovement：lowfrequencysub—beltfailedtotakefmladvantageofcontainsimagesoftheoriginalenergy：itCannotfullyutilizewaveletcoe伍cientchara

8、cteristics；codingtimelongresultingincodingefficiencydecline，DispartFrequencyEmbeddedZerotreeWaveletwasproposed(DFEZW、RecommendMulti—ResolutionmotionEstimationandCompensate(MRME)inwaveletfieldithavemanydisadvantageincludi

9、ngthesportvectordiscontinous，truebodyinmotionborderwithblockborderinconsistent，makethehigh—frequencyweightoftransformcoefficientincrease，reducethecodeefficiencyofDFD，SOthistextproposeMultiResolutionmotionEstimationandCom

10、pensatebasedondiffereneeassignneighburlumppredict(FFMRME)Finallyaccordingtotheimagecoefficientofwaveletdistributioncharacteristic，andthedifferentsensifivityofhumanvisualsystemonedge，smoothregionandtextureoftheimagewavele

11、timageswillbepresentedthathighfrequencypartitionintoedge，smooth，texturezonefractalcoding，throughexperimentsindicatethatthemethodcanbefurtherimprovedimagequalityrestorationKeyWords：Videocoding，wavelettransform，Humanvision