一种基于神经网络、DWT和RSA的医学图像增强水印算法(IJMECS-V8-N4-6)

I.J.ModernEducationandComputerScience,2016,4,46-56PublishedOnlineApril2016inMECS()DOI:10.5815/ijmecs.2016.04.06Copyright©2016MECSI.J.ModernEducationandComputerScience,2016,4,46-56AnEnhancedDigitalImageWatermarkingSchemeforMedicalImagesusingNeuralNetwork,DWTandRSA1SujataNagpal1ChandigarhEngineeringCollege,Landran,Mohali,140307,IndiaEmail:[e-mail]2ShashiBhushanand3ManishMahajan2Professor,DepartmentofCSE,ChandigarhEngineeringCollege,Landran,Mohali3AssociateProfessor,DepartmentofCSE,ChandigarhEngineeringCollege,Landran,MohaliEmail:[e-mail],[e-mail]—Imagewatermarkingistheprocessofthehidingtheoneimageintootherimageforthecopyrightprotection.TheprocessofwatermarkingmustbedoneinthiswaythatthepixelsoftheoriginalimagemustremaininitsoriginalHDform.Alotofworkhasbeendoneinthiscontextinpreviousyearsbutsometechniqueshavetheirownapplications,drawbacksaswellasadvantages.So,thispaperwillutilizethreetechniquesi.e.DiscreetWaveletTransform(DWT),NeuralNetwork(NN)andRSAencryptionforimagewatermarking.IntheendtheperformanceoftheproposedtechniquewillbemeasuredonthebasisofPSNR,MSE,BCR,BERandNCCinMATLABR2010aenvironment.IndexTerms—DigitalImagewatermarking,DWT,NeuralNetwork,Encryption,RSA.I.INTRODUCTIONDigitalformproductslikevideo,audio,textetc.arebeingtransmittedineverydayofourlives[1,2].Inthefieldofdigitalmultimediacommunication,sometechniqueshashelpedalotforstoring,editingandaccessingoftheseproducts[3].But,securityduringthetransferringcommunicationisverycrucial.Aslackofsecurityleadstothelossinpropertyrights.Sotosolvethisproblem,digitalwatermarkingisverygood[4].Fig.1showstheprogressionofwatermarkembedding.Digitalwatermarkingistheprocessofembeddinganimagewithsecretdataforthecommunication.Theembeddedimagecanonlybeextractedbypersonwhohasauthentication.Fig.1.WatermarkembeddingprocessBasicallytherearetwomethodsforwatermarking.i.e.spatialandtransformdomain[5,6].Eachtechniquehasitsownadvantagesaswellasdisadvantages.Butthemainadvantageoftransformdomainmethodisthatitismorerobust,soinproposedworktransformdomainbasedmethodwillbeapplied[7].Fig.2showsthewatermarkextractionprocess.Fig.2.WatermarkextractionprocessTherearetwotypesofimagewatermarking,oneisblindandtheotheroneisnon-blind.Inblindimagewatermarking,thereisnorequirementoforiginalcontentanditisalsocalledaspublicimagewatermarking.Innon-blindimagewatermarking,thereisrequirementoforiginalcontentanditisalsoknownasprivateimagewatermarking.A.MedicalImagesTheamountofdigitalimagesintheinternethasbeenincreasingdaybyday.Theneedofsecurediagnosisinthemedicalfieldisverynecessary.Asthesedaystransmissionofimageshasbecomecommon,sosecurityisneededathighrate.Medicalimagesareusedtoviewtissues,organsorbodypartsofaperson.Theseimagesareusedfordiseasediagnosisandtreatmentofthedisease.Medicalimagesplayamajorroleintheclassificationofdiseases.Thesearealsohelpfulinmonitoringdiseaseactivity.Severalmachinesareaccessibletodaythatcangenerateimagesofinternalorgansofaperson.Therearedifferenttypesofimagingmethodssuchasx-ray,ultrasound,MRIscans,CTscans,Nuclearmedicinescansetc[16].Electromagneticwavesarealsocalledx-rays.X-raysareAnEnhancedDigitalImageWatermarkingSchemeforMedicalImagesusingNeuralNetwork,DWTandRSA47Copyright©2016MECSI.J.ModernEducationandComputerScience,2016,4,46-56thecommonformofmedicalimaging.X-raysarepassedthroughthebodyofapersontoproduceimagesofinternalstructureofbody.X-raysarecommonlyusedforcheckingswallowedobjects,bloodvessels,brokenbonesetc.Ultrasoundisalsoknownasmedicalsonography.Soundwavesareusedinthismethodtocheckinternalorgans.Itisusedinpregnancyortocheckorgansinabdomenandpelvisetc.MRIstandsforMagneticResonanceImaging.Radiowavesandmagneticfieldareusedinthismethodsothatorgansandtissuesofthebodycanviewedinmuchmoredetail.Itisusefultocheckspinalinjuries,abnormaltissues,bloodvesselsetc.CTscansaresimilartoX-raysbutinthistechnique,multipleX-raysareprojectedfromdifferentanglestoobtaina3-Dviewofdifferentorgans.Medicalimagesmustbetakencareoftoconfirmtheidentityasalldiagnosedependsontheobtainedmedicalimage.Inordertofulfillthesecurityandconvenienceissuesofthepatientswatermarkingisusedinthisfield.B.NoiseandDistortionsImagesarevulnerabletovarioustypesofnoises.Noiseisanydegradationinthesignalofanimage.Itiscausedduetoexternaldisturbanceduringtransmissionofanimagefromoneplacetoanother.Differenttypesofdeviceslikenetworkcable,satellitesetc.canbeusedfortransmissionofimage.TherearedifferenttypesofnoisethatcanoccurduringtransmissionsuchasGaussian,Poisson,Salt&Pepperetc.Indigitalimages,Gaussiannoisecanbearousedduringattainmentorcommunication.Thisnoiseisadditiveintemperament.Additivemeansthateachpixelinthenoisedimageistheadditionoforiginalpixelandsomerandomvalue.GaussiannoiseisthetypeofthenoiseinwhichPDFisequaltothenormaldistribution.SometimesitisalsoknownastheGaussiandistributionnoise.ThemodifiedvaluesofGaussiandistributionleadtothistypeofnoise.ThevarioussourcesofGaussiannoisearesensornoise,hightemperatureandelectroniccircuitnoise.Poissonnois