WaveletImageInpaintingBasedon DictionaryLearningwithaBetaProcess Abstract -Theproblemofimageinpaintingandwaveletimageinpainting we representedinthis study . D ictionary L earningwith aB eta P rocess(BPDL) wa sintroduced. AnewmethodbasedonBPDL wa sproposedforwaveletimageinpainting.Unlikeconventionalmethodswhichmostlybasedondiffusion theoryinphysics,thismethodisbasedonsparseimagerepresentationand considers an imageasa combinationofdifferent structuralpatterns toachieveinpainting.Theimage simulat ionexperiments we redesignedtotestthealgorithm.Theresultsdemonstrate dthat theconnectivityprincipleofhuman perception wa swellrealizedwithgoodvisioneffect. ThePSNRofthe waveletcoefficientspartly damaged imageswasimproveds i gnificantly after processed bythenewmethod . It'salsoavailableforNIRimages. It ' s concludedthatthepresentedmethodbasedonBPDLwas aneffectivemethodforwaveletimage inpainting. Keywords - ImageI npainting ,W avelet ,DictionaryLearning,Betaprocess pdf全文下载