Abstract. This paper proposes a new user interface paradigm for motion capture based animation systems, providing intuitive and efficient ways to visualize the main motion capture concepts and operations. A prototype system was built, implementing the prop
Normally,thecaptureddataconsistsofmarker’spositionaland/orangularvariation,sampledbythecapturehardwareduringthenumberofframesre-quiredtocompletetheactor’sperformance.Thisdescriptionisusedineachdegreeoffreedom(DOF)oftheactor.
Internally,theinteractionwiththeuseriscontrolledbyadynamicdatastructurethatrepresentsthecurrent“state”ofallwindowsandmaindatastructuresexistinginthesystem.
2.2AnArchitectureFocusedonMotionCapture
Wedevelopedaconceptualarchitecture,designedtoworkwiththemotioncap-tureparadigm.Itfocusesonsometechnologicalaspectsandembodiesseveraltechniquestodealwithcaptureddata,thusallowingthecreationofreusablemotionlibrariesbyusingabuildingblockparadigm.
Theframeworkofthearchitecture(Figure3)isformedbythreebasicmodules(input,processingandoutput),eachoneresponsibleforaspeci csetoftasks.Thedatastructuresusedinthesystemweredescribedin2.1.
Figure3:System’sarchitecture.
Allmodulesaresupportedbytheconceptualinterfacethatwillbedescribedinsection3.Foramoredetaileddescriptionofsystem’sarchitecture,pleasereferto[4].
InputModule
Theinputmodulefocusesonproblemsconcerningtheinterpretationandpre-processingofmotiondata.Skeletonde nition leswerecreatedtoestablishrelationshipsbetweentheskeletonexpectedbytheincomingdata1andarchi-tecture’sinternaldefaultskeletonde nition.Also,geometricalalgorithmsfor1Dependingonthemotiondataformat.