融合EfficientDet检测算法与虚拟现实技术的儿童护理模拟教学研究

打开文本图片集

AbstractObjectiveToinvestigatethffectivenessofchildcaresimulationteachngapproachthatitegratestheEficientDetdetection algorithmwithvirtualrealitytechnology.Methods：Avirtualhuman-basedteachingplatformwasdevelopedusingvirtualreality technologyTheEfiientDetdetectionalgorithmasenhancedthroughtheincorporationofaspatialatentionmechanismandLaplacian pyramiddecompositiontoenablemultiategoryactiondetection.Furthermore，anitegratedalgoritmcombiningtheediesioal convolutionalneural networks（3D-CNN）with bidirectionallongshort-termmemory（Bi-LSTM）networks wasemployedtocapture spatiotemporalfeaturesfromvideosequencesandachieveaccuraterecognitionofpediatricnursingactions.Results：Theimproved EfficientDet algorithm achieved a mean accuracy of 98.42% ，processing up to 25.14 video frames per second.In the task of nursing action recognition，the3D-CNN-Bi-LSTMmodelattinedaverage accuracy rate，recallrate，andF1 scoresof 95.24% 91.36% ，and 93.68% ， respectivelyutperforingcomparativealgorits.Nurses trainedusingthevtalumanteachingplatfordemonstratedigicantly beterlearningoutcomescompared tothosereceiving traditionalteaching.Conclusion：Theproposednursingactiondetectionand recognitinalgorithsexhibithighprecisionandrelability.Thesimulationteachingframeworkasedonteseagoritsandvirtualreality technologyefectivelyeanceurseseangpfomanceinpdiatricnursingoferaoveandpromsingsotionforrsingucatio.

keywordspediatrics;virtual reality;teaching;nursing education;EfficientDet algorithm

摘要目的：探索融合EfcientDet检测算法与虚拟现实技术的儿童护理模拟教学研究效果。（剩余16232字）