基于小波的圖像壓縮方法及其在視頻壓縮中的應(yīng)用.pdf

1、分類號(hào)UDC密級(jí)學(xué)號(hào)彳癃大季碩士學(xué)位論文論文題目：基于小波的圖像壓縮方法及其在視頻壓縮中的應(yīng)用論文作者：董世都蓑暴、整囂囂蓉：楊小帆教授重慶大學(xué)申請學(xué)位級(jí)別：碩士專業(yè)名稱：計(jì)算機(jī)軟件與理論論文提交日期：2003年5月8Et答辯日期：2003年6月4日學(xué)位授予單位：重慶大學(xué)授位日期：2003年6月30日答辯委員會(huì)主席：朱慶生教授論文評(píng)詞人：鄧亞平教授廖曉峰教授2003年5月8日重慶大學(xué)碩士學(xué)位論文英文摘要ABSTRACTImagecomp

2、ressionplaysailimportantroleintransmissionandstorageofmultimediaWavelethasbeenwidelyadoptedinimagecompressionduetoitsmultiresolutiontime／frequencylocalizationlowertimecomplexityetcWiththeadvanceofImagecompressiontechniqu

3、esbasedonwavelet，avarietywaveletcoefficientmodelsarepresentedSUchasEZWSPⅢTandEBCOTusedinJPEG2000TheyallBleSNRsealable，and，especiallyEBCOTisresolutionscalablen埠sepropertiesRleessentialintransmissionofimageonInternetBecaus

4、etheperformanceofanentropycoderCanbesignificantlyimprovedbyhavingthecoderdynamicallyadapttothecurrent”context”，contextmodelsalsohavewidelybeenadoptedinimagecompressionUsingContextmodel1PEGLSandCALICCanimprovetheircompres

5、sionrateandgainhigherlosslesscompressionratethanEZWSP莊ITandEBCOTInspiredbyGlicbawlsCALIC，ECECOWandEZWcodingschemes，anewcontextmodelofwaveletcoefficientsforimagecompressionisproposedWaveletcoefficientsareencodedbythearith

6、meticencoderwiththecontextsbeingformedbyquantizinglinearpredictionvaluesExperimentalresultsshowthatthemodelachieveshigherlosslesscompressionrateofimagethanlosslessSPIHTandlosslessEBCOTusedinJPEG2000Inadditionbyexploiting

7、themulfiresolufionpropertyofwavelet，themodelcandealwiththe拄ansformedimagefortransmissionpurposeswhichisresolutionscalableandeamhighercompressionrateforeachscaleoftheimagethanEBCOTSomecoefficientmodelsofwaveletforimagecom

8、pressionhavealreadybeenextendedtovideocasesusingThreeDimensionalwaveletIntheSalllewayaftermodificationPCWisalsoextendedto3D—PCWwhichisresolutionscalableintimeandspaceFirstthecoefficientsofwaveletarequantizedbythethreshol

9、dformedbytheeye’SsensitivitytoitssubbandThenWaveletCOefficientsareencodedbytheentropyencoderwiththecontextsformedbyquantizingtheirprediorionvalues，whicharedeterminedbytheirneighborshispaceandtimedimensionsSince3D—PCWtack