简介:Weconsiderforafixedμ,theclassofpolynomialsPn,μ,s:=P(z)=zsanzn-s+n-s∑j=μan-jzn-j-s;1≤μ≤n-sofdegreen,havingallzerosin|z|≤k,k≤1,withs-foldzerosattheorigin.Inthispaper,wehaveobtainedinequalitiesinthereversedirectionfortheaboveclassofpolynomials.Besides,extensionsofsomeTuran-typeinequalitiesforthepolarderivativeofpolynomialshavebeenconsidered.
简介:Theworldisundergoingconstantandcomplicatedchange.Development,adjustmentandtransformationmakeuptheworld’smainmelody,whilepeaceanddevelopmentarestillthecurrentthemes.Multi-polarizationandeconomicglobalizationcontinuetobecomemoreentrenched,andculturaldiversityandsocialnetworkingacross
简介:Motiondeblurringisabasicprobleminthefieldofimageprocessingandanalysis.Thispaperproposesanewmethodofsingleimageblinddeblurringwhichcanbesignificanttokernelestimationandnon-blinddeconvolution.Experimentsshowthatthedetailsoftheimagedestroythestructureofthekernel,especiallywhentheblurkernelislarge.SoweextracttheimagestructurewithsalientedgesbythemethodbasedonRTV.Inaddition,thetraditionalmethodformotionblurkernelestimationbasedonsparsepriorsisconducivetogainasparseblurkernel.Butthesepriorsdonotensurethecontinuityofblurkernelandsometimesinducenoisyestimatedresults.ThereforeweproposethekernelrefinementmethodbasedonL0toovercometheaboveshortcomings.Intermsofnon-blinddeconvolutionweadopttheL1/L2regularizationterm.Comparedwiththetraditionalmethod,themethodbasedonL1/L2normhasbetteradaptabilitytoimagestructure,andtheconstructedenergyfunctionalcanbetterdescribethesharpimage.Forthismodel,aneffectivealgorithmispresentedbasedonalternatingminimizationalgorithm.