简介:Inthispaper,analgorithmthatdeterminesarealalgebraiccurveisoutlined.Itsbasicstepistodividetheplaneintosubdomain1sthatincludeonlysimplebranchesofthealgebraiccurvewithoutsingularpoints.Eachofthebranchesisthenstablyandefficientlytracedintheparticularsubdomain.Exceptfortracing,thealgorithmrequiresonlyacoupleofsimpleoperationsonpoly-nomialsthatranbecarriedoutexacrlyifthecoefficientsarerational,andthedeterminationoftherealrootsofseveralunivariatepolynomials.
简介:Apiecewisecubiccurvefittingalgorithmpreservingmonotonicityofthedatawithoutmodificationoftheassignedslopesigproposed.ThealgorithmhasthesameorderofconvergenceasYan’salgorithm[8]andGasparo-Morandi’salgorithm[5]foraccurateorO(hq)accurategivendata,butithasamorevisuallypleasingcurvethanthosetwoalgorithms.WealsodiscusstheconvergenceorderofcubicrationalinterpolationforO(hq)accuratedata.
简介:Itisknownfromclassicaldifferentialgeometrythatonecanreconstructacurvewith(n-1)prescribedcurvaturefunctions,ifthesefunctionscanbedifferentiatedacertainnumberoftimesintheusualsenseandifthefirst(n-2)functionsarestrictlypositive.ItisestablishedherethatthisresultstillholdsundertheassumptionthatthecurvaturefunctionsbelongtosomeSobolevspaces,byusingthenotionofderivativeinthedistributionalsense.ItisalsoshownthatthemappingwhichassociateswithsuchprescribedcurvaturefunctionsthereconstructedcurveisofclassC∞.
简介:Theappriximationpropertiesofgeneralizedconiccurvesarestudiedinthispaper.Ageneralizedconiccurveisdefinedasoneofthefollowingcurvesortheiraffineandtranslatione-quivalentcurves:(i)coniccurvesiincludingparabolas,hyperbolasandellipses;(ii)generalizedmonomialcurves,includingcurvesoftheformx=yr,.rR.r=0,1,inthex-yCartesiancoordinatesystem;(iii)exponentialspiralcurvesoftheformp=Apolarcoordinatesystem.Thistypeofcurveshasmanyimportantpropertiessuchasconvexity,approximationproperty,effectivenumericalcomputationpropertyandthesubdivisionpropertyetc.Applicationsofthesecurvesinbothinterpolationandapproximationsusingpiecewisegeneralizedconicsegmentarealsodeveloped.Itisshownthatthesegeneralizedconicsplinesareverysimilartothecubicpolynomialsplinesandthebesterrorofapproximationisoratleastingeneralprovidedappropriateproceduresareused.Finallysomenumericalexamplesofinterpolationandappro
简介:Receiveroperatingcharacteristic(ROC)curvesareoftenusedtostudythetwosampleprobleminmedicalstudies.However,mostdatainmedicalstudiesarecensored.UsuallyanaturalestimatorisbasedontheKaplan-Meierestimator.InthispaperweproposeasmoothedestimatorbasedonkerneltechniquesfortheROCcurvewithcensoreddata.Thelargesamplepropertiesofthesmoothedestimatorareestablished.Moreover,deficiencyisconsideredinordertocomparetheproposedsmoothedestimatoroftheROCcurvewiththeempiricalonebasedonKaplan-Meierestimator.ItisshownthatthesmoothedestimatoroutperformsthedirectempiricalestimatorbasedontheKaplan-Meierestimatorunderthecriterionofdeficiency.Asimulationstudyisalsoconductedandarealdataisanalyzed.
简介:全部的变化(电视)最小化问题广泛地在图象恢复被学习。尽管许多其他的方法为它的答案被建议了,牛顿方法由于没有集中为最初的明确的表达仍然保持不可用。由禅宗,周和禅宗的以前的研究[15]认为一个规则化参数继续想法与一些成功而是没有柔韧的参数选择增加牛顿方法的集中的领域计划。在这份报纸,我们为一样的最初的电视明确的表达考虑一个homotopy方法并且建议使用曲线追踪适应地选择规则化参数。结果,这个想法帮助实质地在高效地解决TVEuler-Lagrange方程改进以前的工作。一样的想法也象deblurring问题一样为二个另外的方法被考虑,再与获得的改进。数字实验证明我们的新方法为图象恢复柔韧、快,甚至为有大noisy-to-signal比率的图象。
简介:精力最小化广泛地被使用了在象电脑辅助的几何设计那样的地里构造曲线和表面,计算机图形。然而,我们的严峻的例子证明精力最小化不有时优化曲线的形状。这份报纸学习在最小化紧张精力和曲线形状之间的关系,学习被与令人满意的形状构造一条立方的Hermite曲线执行。立方的Hermite曲线插入内推二个给定的端点的位置和正切向量。计算机模拟技术成为了科学发现的方法之一,学习进程被数字计算和计算机模拟技术执行。我们的结果显示出那:(1)立方的Hermite曲线不能被完全最小化紧张精力构造;(2)紧张精力由本地最小的采纳珍视,立方的Hermite曲线的形状能为大约60%所有情况被决定,其中一些然而有不能令人满意的形状。基于种类精力模型和分析,一个新模型为与令人满意的形状构造立方的Hermite曲线被介绍,它是种类精力的修正模型。新模型使用一个明确的公式计算二正切向量的大小,并且有性质:(1)计算是容易的;(2)它让立方的Hermite曲线当保持在曲线建设为一些盒子最小化种类精力的好性质时,有令人满意的形状。与最小的种类精力模型一起的新模型的比较被包括。
简介:AnefficientmethodforC2nearlyarc-lengthparameterizedcurveispresented.Anideaofapproximationforthearc-lengthfunctionofparametriccurvewhichinterpolatesCADdatapointsisdiscussed.Theparameterizationisimplementedbyusingparametertransformation.Finally,twonumericalexamplesaregiven.