基于JPEG圖像的隱寫算法研究.pdf

1、南京理工大學(xué)碩士學(xué)位論文基于JPEG圖像的隱寫算法研究姓名：賀麗芳申請學(xué)位級別：碩士專業(yè)：檢測技術(shù)與自動化裝置指導(dǎo)教師：劉光杰20100620碩士論文基于JPEG圖像的隱寫算法研究AbstractInformationhidingisanimportantissueinthefieldofcurrentinformationsecurityAsthemainbranchofinformationhiding，steganographyh

2、asdrawnmoreandmoreattentionfromextensiveresearchersWiththedevelopmentofsteganographysteganalysisusedtodetecttheexistenceofsecretmessageinimages，hasmaderapiddevelopment，andofthebothmutuallypromoteeachotherandachievethecom

3、mondevelopmentForJPEGimages，basedonthesummaryofpositiveandnegativetechnologiesinthefield，adigitalimagesteganographicalgorithmresistingthecurrentefficientsteganalyticalgorithmsisproposedandanalyedThemainworkincludes：(1)Th

4、efourcommonsteganographicalgorithmsbasedonJPEGimagesarestudied，andJSteg，F(xiàn)5andMBsteganographicalgorithmsarecomparedandanalyzedaccordingtotheembeddingcapacityembeddingrate，datautilizationrate，datachangingrate，PSNR，statisti

5、calpropertiesofDCTcoefficientsandSOonviaexperiments，andTheadvantagesanddisadvantagesofeachalgorithmareanalyzed(2)ThethreemoretypicaluniversalsteganalysisalgorithmsbasedonJPEGimagesarestudied，andareevaluatedbyJSteg，F(xiàn)5，Out

6、GuessandMBsteganographicalgorithmsmentionedaboveTheexperimentalresultsshowthatthedetectionratesofthethreesteganalyticalgorithmsaremorethan90％underfullyembedding，thedetectionreliabilitylookniceAnd，resultsalsoshowthatthepe

7、rformanceofthesteganalyticalgorithmsarecloselyrelatedtothesensitivityoffeaturesfacingsteganographicoperations(3)BasedonthestatisticalcorrelationsofDCTcoefficientsofintrablockandinterblock，AsteganographicalgorithmbasedonM

8、arkovchainisproposedOnesteptransitionprobabilitymatrixoftheMarkovchainandpseudo—randomsequenceareusedtodeterminethemodificationruleoftheDCTcoefficientsintheLSBmatching，whichminimizestheimpactofthecorrelationofimageblocks

9、comingfromdataembeddingTheexperimentsalsoshowthatcomparedwithothersteganographicalgorithms，itscapacityislargerthedatautilizationrateishigheritsPSNRvalueincreases，andthestatisticspropertiesofDCTcoefficienthistogramarepres

10、ervedComparedtoseveralothersteganographicalgorithms，thecapabilityagainstthesteganalyticalgorithmsisimprovedtosomeextentFinallytheexistingdeficienciesofthedissertationarepointedout，andthefurtherresearchworksarepresentedII