融合局部感知增强的投篮上肢动作分解与高精度识别算法

  • 打印
  • 收藏
收藏成功


打开文本图片集

关键词:HRNet模型;双流金字塔;图卷积网络;时间卷积网络;人体动作识别;图像分析中图分类号:TN911.73-34;TP391 文献标识码:A 文章编号:1004-373X(2025)18-0119-06

Shooting upper limb motion decomposition and high-precision recognition algorithm integrating local perception enhancement

RENYufei,LIUHailin² (1.Shaanxi College ofCommunications Technology,Xi'an710o18,China; 2.Shaanxi Transportation ElectronicEngineering TechnologyCo.,Ltd.,Xi’an71Oo68,China)

Abstract:Inalusion tothe problems of jointoclusionandlarge receptive field existing in traditional motion recognition algorithmswhenprocesing basketballtraining images,ahigh-precisionupperlimbmotiondecompositionrecognitionmodel integratinglocalperceptionenhancementisproposed.Themodelcanextracttheglobalstructuralfeaturesofthehumanbodyby meansof lightweighthigh-resolution network (HRNet)branches,andcombinewithatwo-streampyramid module toenhancelocal jointperceptionability.Inthespaceflow,thecavityconvolutionisusedtoexpandthehandreceptivefield,andinthetieflow, theopticalflowisusedtocapturetherelativemotionofthegolfer,significantlyimprovingthehandpositioningaccuracyinthe occluded scene.Atthesametime,thespatiotemporaldecompositionmoduleisalsodesignedinthemodel toaalyzethespatial andtemporalcharacteristics,constrainthenetworkfocusandlearnthemotionconsistencybetwen frames,andmakingthedata procesing processmorerobust.Theadaptiveweighting elementsare introduced intothemodeltooutput high-precision joint pointcoordinates.The experimental testing results show that themAP @0.5 ofthe proposed model can reach 86.9%, the F1 value of action classification is 95.3% ,whichisbetter than the current mainstream algorithm.and thekeyframe detectionrateis 91.1% ,theframeratecanreach47f/s,andthereal-timeperformanceisgood,whichfullproves thatthe modelcanprovide reliable technical means for basketball training and intelligent analysis of events.

Keywords:HRNet model; two stream pyramid; graph convolutional network;time convolutional network;human motion recognition; image analysis

0 引言

在篮球运动中,完美的投篮动作取决于压腕角度、手指对球的控制以及肘关节在篮球出手瞬间的伸展角度与轨迹。(剩余6516字)

monitor
客服机器人