摘要
Anewlocallyadaptiveimagedenoisingmethod,whichexploitstheintrarscaleandinter-scaledependencyinthedual-treecomplexwaveletdomain,ispresented.Firstly,arecentlyemergedbivariateshrinkageruleisextendedtoacomplexcoefficientanditsneighborhood,thecorrespondingnonlinearthresholdfunctionsarederivedfromthemodelsusingBayesianestimationtheory.Secondly,anadaptiveweight,whichisabletocapturetheinter-scaledependencyofthecomplexwaveletcoefficients,iscombinedtotheobtainedbishrinkthreshold.Theexperimentalresultsdemonstrateanimproveddenoisingperformanceoverrelatedearliertechniquesbothinpeaksignal-to-noiseratio(PSNR)andvisualeffect.
出版日期
2007年03月13日(中国Betway体育网页登陆平台首次上网日期,不代表论文的发表时间)