A STATISTICAL WATERMARKING SCHEME FOR 3D MESH MODELS BASED ON VERTEX SMOOTHNESS MEASURE

Abstract

3D imesh iwatermarking ihas isparked iinterest, ithanks ito ithe isudden iincrease iof igraphic icontent iand ianimated imotion imovies, ihowever ithe istate iof ianalysis iof iwatermarking i3D imodels icontinues ito ibe iin iits iinfancy ias icompared ito iresearch iwork iin iimage iand ivideo iwatermarking. i3D iWatermarking iprovides ia ideterrent ito ipiracy iof i3D imodels iby iembedding ia ihidden ipiece iof idata iwithin ithe ioriginal icontent. iWatermarking ialgorithms ihave ia ibasic idemand ithat ithe iwatermark iought ito ibe iimperceptible ito iavoid ibeing idetected iand inot icause ivisible idistortion ito ithe iviewer. iThe iwatermark iought ito ieven ibe isturdy ito iface iup ito iunintentional iattacks. iIt's iconjointly idesired ithat ithe iwatermark iinsertion icapability iought ito ibe ias ihigh ias iattainable ito iface iup ito iintentional iattacks iand ito ipermit iinsertion iof imultiple ior iredundant ior ibio-metric iwatermarks. iInsertion iof ihigh idensity iimperceptible iwatermark ican icreate iit ivery itough ifor ian iattacker ito iseek iout ithe iwatermark iand iso icreate isubstantial ichanges iwithin ithe i3D imodel ito iget irid iof ior ichange ithe iwatermark. iHowever, iinserting ihuge iamounts iof idata ias iwatermark iwill icause idistortion; itherefore idesigning iof iwatermarking ialgorithms iinvolves ia itrade-off ibetween iimperceptibility, icapacity iand irobustness. This ithesis idiscusses iblind iwatermarking iof i3D imodels iwithin ithe ispatial idomain. iAlthough iit's ibeen inotable ithat ioblivious i(or iblind) iwatermarking ischemes iare iless isturdy iand irobust ithan inon ioblivious iones, ithey're ihelpful ifor inumerous iapplications iwherever ia ihost isignal iisn't iout ithere iwithin ithe iwatermark idetection iprocedure. iThe itarget iof ithis iresearch iwork iis ito iexplore iinnovative iways iin iwhich ito iinsert ithe imost iquantity iof isecret iinfo iinto i3D imesh imodels iwhile inot iinflicting ipalpable idistortion iand iconjointly icreate iit itough ifor ithe iattacker ito iguess iwhere ithe iwatermark iwas iinserted iand itherefore ithe iquantity iof iwatermark iinserted. The iproposed imethodology iestimates ithe ilocal ismoothness ivariation iof ithe imesh ito ipick ivertices ifor iinserting ia iwatermark. iSmoothness ivariation iof ithe isurface irepresented iby ithe i1- iring ineighborhood iof ievery ivertex iis icomputed iby ithe iaverage iangle idifference ibetween v

ithe isurface inormal iand ithe iaverage inormal. iEvery ivertex iof ithe imesh iis ilabeled iin ione iof ithe ibins icorresponding ito ivariable idegrees iof ilocal ismoothness ivariation ifrom ithe ilow ito ithe imoderate ito ithe ihighest ivariation. iVertices iwith ithe ilabel iof imoderate ilocal ismoothness ivariation iare ithen ipicked ifor ithe iinsertion iof ia irandom iwatermark. iThe iprojected imethodology iemploys ia istatistical iapproach ito iembed iwatermark ion iselected ivertices ithat ichanged ithe idistribution iof ivertex inorms iconsistent iwith ithe iwatermark ibit ito ibe iembedded. iHistogram imapping ifunctions iare iintroduced ito imodify ithe idistribution ias iper ithe iwatermark irequirement. iThese imapping ifunctions iare idevised ito ilessen ithe ivisibility iof iwatermark ias imuch ias ipossible. iSince ithe istatistical ifeatures iof ivertex inorms iare iinvariant ito ithe idistortion-less iattacks, ithe iproposed imethod iis isturdy iagainst irotation, itranslation iand iscaling. iMoreover, iour imethod iuse ian ioblivious iwatermark idetection ischeme, iwhich imay iextract ithe iwatermark iwhile inot irequiring ithe icover imesh imodel. iSimulation iresults iprove ithat ithe iinserted iwatermark iis iinvariant iagainst iaffine ioperations, inoise iand ismoothing iattacks, iand iat ithe isame itime iimperceptibility iis iretained. i The ithesis iis istructured ias ifollows: i Chapter-1 gives an introduction about watermarking, its categories, requirements and applications, also giving an overview of the trade-off involved in designing desired watermarking schemes. Chapter-2 gives a background of 3D mesh entities and basic terminologies used in thesis Chapter-3 gives a comprehensive literature review of the contemporary 3D watermarking algorithms. Chapter-4 describes proposed watermarking methods in detail, including their insertion and embedding algorithm based on statistical features and vertex smoothness measure of a 3D mesh. Chapter-5 shows the experimental results and analysis of the proposed method against various attacks. Chapter-6 gives the conclusion and the future work of the thesis.