利用变换的高效图像隐写算法：Wavelet和Contourlet与BooFISH加密(IJCNIS-

I.J.ComputerNetworkandInformationSecurity,2015,2,15-24PublishedOnlineJanuary2015inMECS()DOI:10.5815/ijcnis.2015.02.02Copyright©2015MECSI.J.ComputerNetworkandInformationSecurity,2015,2,15-24EfficientImageSteganogrphicAlgorithmsUtilizingTransforms:WaveletandContourletwithBlowfishEncryptionSaddafRubabDepartmentofComputerEngineering,CollegeofElectrical&MechanicalEngineering,NationalUniversityofSciences&Technology(NUST),Islamabad,[e-mail]omputerEngineering,CollegeofElectrical&MechanicalEngineering,NationalUniversityofSciences&Technology(NUST),Islamabad,[e-mail]—Steganographyisameanstohidetheexistenceofinformationexchange.Usingthistechniquethesenderembedsthesecretinformationinsomeothermedia.Thisisdonebyreplacinguselessdatainordinarycomputerfileswithsomeothersecretinformation.Thesecretinformationcouldbesimpletext,encodedtextorimages.Themediausedastheembeddingplanecouldbeanimage,audio,videoortextfiles.Usingsteganographyensuresthatnooneapartfromthesenderandthereceiverknowsabouttheexistenceofthemessage.Inthispaper,asteganographymethodbasedontransformsusedi.e.WaveletandContourlet.Devisedalgorithmwasusedagainsteachtransform.BlowfishEncryptionmethodisalsoembeddedtodoublethesecurityimpact.Themajoradvantageofapplyingtransformsisthattheimagequalityisnotdegradedevenifthenumberofembeddedcharactersisincreased.Theproposedsystemoperateswellinmostofthetestcases.Theaveragepayloadcapacityisalsoconsiderablyhigh.IndexTerms—Steganography,PSNR,blowfish,Huffman,DCT,DWT.I.INTRODUCTIONTheartofhidinginformationinsomeothermediaofinformationisknownassteganography.Thesecretinformationisembeddedintoanothermediainsuchawaythatnooneelsethansenderandpropspectiverecievercanknowaboutthepresenceofthesecretinformation.Inotherwords,usingthistechniqueevenobscurethatasecretinformationisbeingtransmitted[1].A.Encodingschemes1.HuffmanInthepastyears,researchersgivemoreattentiontovariablelengthcodesamongthenumerousencryptionalgorithmsincludingDES[2],AES[3],andTripleDES[2].HuffmancodeisamappingmethodforsymbolstobitsofvariablelengthlikearithmeticcodingandLempel-Zivcoding.DavidA.Huffmanin1952[4],aPhDstudentatMITintroducedthisencodingschememainlyforthepurposeoflosslessdatacompression.Itgivesmemorymanagementbyusingfewlocations/bits.Asanintermediateoutputitgeneratesabinarytreeusingsymbolsandsometimesletterswiththeirrespectivefrequenciesinascendingordescendingorder.Thesumofsmallestfrequenciesisassignedtoparentnode,andthiscontinuesuntilweareleftwithonlytherootnode.Totranslatethecodes,0‘sand1‘s(binarydigits)areassignedtoallnodesandreadfromleaftorootnode.TheoutputistabulatedinHuffmanTable.Huffmancodesprovidesbettersafeguardtosecretdatafromanythirdpartyattacks.2.BlowfishBruceSchneier,acryptologist,designeda64-bitblockcipherandvariablelengthkeyandproducedwiththenameBlowfish.Theprocessisdividedintwopartsi.e.theexpansionofkeyandencryptionofdata.Itiswidelyusedtoreducetherepeatinginformationresultinginincreasedcontrastimage.BlowfishwillbeusedinourproposedmethodfordataencryptionbecauseitsperformanceoutshinesifparalleledwithotherencryptionalgorithmslikeAES,DES[1].16EfficientImageSteganogrphicAlgorithmsUtilizingTransforms:WaveletandContourletwithBlowfishEncryptionCopyright©2015MECSI.J.ComputerNetworkandInformationSecurity,2015,2,15-24Fig1:Short-TimeFourierTransformofaSignalII.BACKGROUNDANDRELATEDWORKHidingdatamethodsinvolvingLeastSignificantBit(LSB)ofimagesaremorepronetodetectionofanysecretdata.SteganographictechniquesproposedinvolvetransformsofimagelikeFourierandDiscreteCosineTransform(DCT),whichmakesthedatatobemoreresistantfrombecomingcorrupted[5].InDCTmethods,itispossibletorecoverdataevenifsomeimagedataislostoralteredbyanythirdparty[8].DCTmodulatesatleasttwocoefficientsofanimageblocktostegantheinformation[6].FourierTransformworkonthesameprinciplesofDCT,bymodifyingcoefficientsofmidbandfrequencies[7].Followingsectionsshallgivethedetaileddescriptionforwaveletandcontourlettransforms:-A.WaveletTransformWavelets‘historyisnottimeworn,butithadagreatinterestofresearcherswhoboostedthewavelettransforminformationintheirarticles.Fouriersetthebasisofthefrequencywhichforalongtimewasthebestandtheonlyapproachexistentinsignalanalysis.Theresearchgraduallymovedfromfrequencybasedanalysistoscalebasedanalysiswhentheresearchersrealizedthatanapproachmeasuringaveragefluctuationsatdifferentscalesmightprovelesssensitivetonoise.Andso,thewavelettransformwasborn.Thefirstrecordedmentionofwhatwecallnowa―wavelet‖datesbackto1909inAlferdHaar‘sthesis[9].TheconceptofwaveletsinitspresenttheoreticalformwasproposedlaterbyMorlet.ThemainalgorithmusingfilterswasprovidedbyStephaneMallat[10]in1989.Sincethen,theresearchhasbecomeinternational.TheMallatalgorithmisinfactaclassicalschemeknowninthesignalprocessingcommunityasatwochannelsubbandcoder[10].Therequirementofvaryingwindowsizewasfulfilledbywaveletanalysiswhichoffersawindowingtechniquewithvariablesizedwindow.Thewavelettransformallowstheuseoflongtimeintervalswherew