Please use this identifier to cite or link to this item: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4921
Full metadata record
DC FieldValueLanguage
dc.contributor.authorJagoda, J. K. K. K.-
dc.date.accessioned2025-08-18T09:16:16Z-
dc.date.available2025-08-18T09:16:16Z-
dc.date.issued2025-06-30-
dc.identifier.urihttps://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4921-
dc.description.abstractAbstract Inrecentyears,CoverlessImageSteganography(CIS)hasemergedasapromis- ing alternativetotraditionalsteganographictechniquesbyeliminatingtheneed forexplicitdataembeddingandinsteadutilizingnaturallyoccurringorgenerated imagesascarriers.However,currentCISmethodsoftensu!erfromlowhiding capacity,limitedtoconcealingasingleimage,ando!er insu”cientrobustnessand securitywhenhandlingmultiplesecretimagessimultaneously. To addresstheselimitations,wepresentanovelmulti-imagecoverlessstegano- graphic frameworktailoredforRGBimages,capableofsecurelytransmittingand reconstructinguptofoursecretimagesusingasinglefusedcarrierimage.Un- likeconventionalmethodsthatdependonpre-existingcoverimages,ourmodel generates anintermediatecarrierimagethroughasophisticatedpixel-levelfusion techniquethatmergesvisualinformationfrommultipleRGBsecretimageswhile ensuringcompleteperceptualinvisibilityoftheoriginalcontent.Thisfusionpro- cess guaranteesthatthecarrierimagecontainsnovisuallyidentifiabletracesof the concealedimages,o!eringaninherentlysecureandstealthyrepresentation. To furtherstrengthenconcealment,thefusedcarrierundergoesmultipleroundsof imagescramblingtransformations,makingunauthorizeddecodingvirtuallydi”- cult.Wealsointegratearobuststeganographickeymechanism,restrictingaccess and decodingcapabilitiesexclusivelytoauthorizedreceiversequippedwiththe correctkey.Thisensuresthatevenifthecarrierisintercepted,thehiddencontent remainssecureandunrecoverablewithouttheproperdecryptioncontext. In thedecodingphase,weproposearefinementalgorithmdesignedtorestore fine-graineddetailslostduringthefusionandscramblingprocesses.Thisalgo- rithmsignificantlyenhancesreconstructionfidelityandvisualquality,eveninthe presenceofadversarialconditionssuchasnoise,tampering,orgeometrictransfor- mations. ComprehensiveexperimentsonbenchmarkRGBimagedatasetsvalidatetheef- fectivenessofourproposedmodel,demonstratingsubstantiallyhigherhidingca- pacity,minimalvisualdistortion,andresilienceagainststatisticalandperceptual steganalysisattacks.ComparedtoexistingCISapproaches,ourmethodo!ers a moresecure,high-capacity,andvisuallyimperceptiblesolutionformulti-image coverlesssteganography,layingthefoundationfornext-generationcovertcommu- nicationsystems.en_US
dc.language.isoenen_US
dc.subjectMulti-imageSteganography,CoverlessImageSteganogra- phy,SteganographicKey.en_US
dc.titleCoverless Multi-Image Steganography for RGB Images: Generative Adversarial Networks based Approach & Steganographic Key based Approachen_US
dc.typeThesisen_US
Appears in Collections:2025

Files in This Item:
File Description SizeFormat 
20000804 - J.K.K.K.Jagoda - Kimuthu Kisal.pdf25.29 MBAdobe PDFView/Open


Items in UCSC Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.