Page 116 - Demo
P. 116


                                    114Chapter6oftheHoloLensstimulatestheusertoinspecthologramsfromdifferentpositionsandviewanglesandthebenefitsof3DarethereforefullyutilizedNeverthelessthe technologyof ARisdevelopingrapidlyanditcanbeexpected thatin thenearfutureupdatesotherdevicesorevenothertechniqueswillcertainlyimprovethecurrentstateofAR TheseinnovationscouldmaketheimpactandapplicationofARinthemedicalfieldevenmoresignificantACCURACYAnaccuracystudyoftheAR-basedIGSsystemwasperformedsincethemostimportantquestionforthefunctionalityofanIGSsystemisiftheaccuracyissufficientfortheintendedapplication37TherequiredaccuracycompletelydependsonthetypeofinterventionyetformostIGSproceduresinoralandcraniomaxillofacialsurgeryneurosurgeryandtraumasurgeryanaccuracyof1to2mmisconsideredacceptable111438-40 Theaccuracyisaffectedbyerrorsofthesystemuser,andvisualizationThenavigationerrorwasdeterminedusingtightfitARexperiments Theresultsoftheseexperimentsshowedthatthemeanerrorofthecompletesetupincludingtheregistrationerror,printingerror,anderrorsintroducedbytheopticaltrackingsystemwas06mminalaboratorysetupByimplementinganewtrackableobjecttheHoloLensasmallincreaseof01mminerrorwasseen(06to07mm)Althoughthisdifferencewasnotsignificantthe01-mmincreaseofthenavigationerrorwasincludedintheerrorevaluationdescribedbelowTheusererrorinthisstudycanbeestimatedwhentheresultsoftheloose-fitIGSexperimentarecompared with theresultsof the tight-fitIGSexperimentsince theonlydifferencebetweentheseexperimentswasthattheuserwasnotrestrictedwhenpositioningthecubeonthemeasuringboard Thetight-fitIGSexperimentshowedameanEuclideandistanceof06mmandtheloose-fitIGSexperimentshowedameanEuclideandistanceof07mmFromtheseresultstheusererrorwasestimated tobe01mmin theloose-fitIGSexperiment Wearing theHoloLensintheloose-fitARexperimentmighthaveaneffectontheusererror,andthiseffectisincludedintheARvisualizationerrorSinceitwasnotpossibletoperformdirectmeasurementsonvirtualobjectswewerenotabletoseparatethiseffectfrom the AR visualizationerror Toreduce theeffectof theusererror toaminimumwemadetheexperimentssimpletoperformfortheobserversThiswasprovedbytheresultingusererrorof01mmintheloose-fitIGSexperimentTherefore
                                
   110   111   112   113   114   115   116   117   118   119   120