Font Size: a A A

Robust Digital Watermarking Algorithms Based On Transform Domain And Singular Value Decomposition

Posted on:2022-03-16Degree:MasterType:Thesis
Country:ChinaCandidate:Y C XuFull Text:PDF
GTID:2518306539995139Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
Therapiddevelopmentofdigitalproductsbringssecurityissues.People haveincreasinglystrictrequirementsonthesecurityandauthenticityof privatecontent.Digitalwatermarkingtechnologyisanimportantmeansof copyrightprotectionfordigitalproducts.Ithidestheidentificationinformation inmultimediadatathroughtheinsensitivityofthehumanvisualperception systemandtheredundancyofdigitalsignals.Basedonavarietyoftransform domainwatermarkingtechniquesandchaoticencryptionmethods,twonew digitalwatermarkingalgorithmsaredesignedinthisdissertation.Themain researchcontentsareasfollows:Toenhancethesecurityofimagewatermarkingalgorithm,multi-chaotic mapsareadoptedtoscramblethewatermarkinginformation.Adoubledigital watermarkingalgorithm basedoncurvelettransform andchaoticmapsis presented.Thealgorithm achievesthattwodigitalwatermarkinginformation are embedded simultaneously in the hostimage.The hostimage is decomposedintosixpartswithcurvelettransform,andarobustwatermarkis embeddedinthecoarselayerforcopyrightprotectionandauthenticationof digitalproducts.Inaddition,anauthenticationwatermarkisembeddedinthe finelayertodetectwhetherthehostimagehasbeentampered.Simulation resultsshow thattheproposeddigitalwatermarkingalgorithm possesses acceptablerobustnessandsecurity.Infunction,itnotonlyrealizesthe protectionofcopyright,butalsorealizesthedetectionofmalicioustampering ofthecarrierimage.Adigitalwatermarkingalgorithm basedoncontourlettransform,discrete wavelettransform,discretecosinetransform,Zigzagscanning,andsingular valuedecompositionisproposed.Inthewatermarkembeddingprocess,the contourlettransform isfirstappliedtoextractthelowfrequencypartofthe hostimage,and then the low frequencypartis processed bywavelet transform.Afterwards,twointermediatefrequencypartsareselectedasthe embeddingareaofthewatermark.Thetwointermediatefrequencypartsare performedadiscretewavelettransform,andthehigh-frequencyresultsare executedbythediscretewavelettransform.Finally,thestabilitycoefficients areextractedthrough Zgzagscanningtoconstructthecarriermatrix.Thei watermarkinformationisembeddedbymodifyingthesingularvalueofthe carriermatrix.Beforethewatermarkinformationisembedded,thechaotic seriesgeneratedby Lgisticmapisusedforscramblingprocessing,henceo improvingtherobustnessofthewatermarkingsystem.Experimentalresults show thatthedigitalwatermarkingschemeisfeasibleandrobustagainst mostcommonattacks.
Keywords/Search Tags:Curvelet transform, Contourlet transform, Discrete wavelet transform, Chaoticsystem, Singularvaluedecomposition
PDF Full Text Request
Related items