简介:Currently,3Dprintingofthecharacterdollsisaverypracticalapplicationfortheaverageperson.Butthemodelofdollwhichcanbeobtainedisstaticsothepostureofthedollissingle.Ontheotherhand,themodificationofthemodelisverydifficulttonon-professions.Thispaperproposesanrapidgenerationmethodofcharacterdollwithrotatablelimbs,whichisthroughaddingthespherejointtothedoll’smodelautomatically.Afterthemodelissegmentedbydrawingalineinteractively,thespherejointiscreatedbasedonthesegmentationboundarythroughentitymodelingmethod.Lastlythetwomodelsofthedollandthejointarecompositedandprinted.Somedoll’smodelaretestedontheFDM(FusedDepositionModeling)3Dprinterusingthisprocess.Theresultsaremoreinterestingandtheefficiencyhasbeengreatlyimprovedcomparedwithmodifyingthemodelmanually.
简介:Mobileplatformisnowwidelyseenasapromisingmultimediaservicewithafavorableusergroupandmarketprospect.Tostudytheinfluenceofmobileterminalmodelsonthequalityofsceneroaming,aparametersettingplatformofmobileterminalmodelsisestablishedtoselecttheparameterselectionandperformanceindexondifferentmobileplatformsinthispaper.Thistestplatformisestablishedbasedonmodeloptimalityprinciple,analyzingtheperformancecurveofmobileterminalsindifferentscenemodelsandthendeducingtheexternalparameterofmodelestablishment.Simulationresultsprovethattheestablishedtestplatformisabletoanalyzetheparameterandperformancematchinglistofamobileterminalmodel.
简介:TheprecisionforgingprocessissimulatedbycommercialsoftwareDeform3Dusingarigidvisco-plasticmodeltopredictthestatusofmetalflowandthedistributionofequivalentplasticstrain,providingguidanceformakingdecisionontheoptimalchoiceofprocessparametersandmouldstructure.TrialforgingwasusedtoverifytheeffectivenessofFEMsimulationresults.
简介:Animprovedself-organizingfeaturemap(SOFM)neuralnetworkispresentedtogeneraterectangularandhexagonallatticwithnormalvectorattachedtoeachvertex.Aftertheneuralnetworkwastrained,thewholescattereddataweredividedintosub-regionswhereclassifiedcorewererepresentedbytheweightvectorsofneuronsattheoutputlayerofneuralnetwork.Theweightvectorsoftheneuronswereusedtoapproximatethedense3-Dscatteredpoints,sothedensescatteredpointscouldbereducedtoareasonablescale,whilethetopologicalfeatureofthewholescatteredpointswereremained.