摘要
本文提出一种基于感知器网络的太极拳关键动作识别算法。首先通过背景差分法对于太极拳运动员目标进行检测。其次设计一个多层感知器网络用于识别太极拳关键动作。最后构建一个太极拳关键动作数据集,根据梯度下降法得到感知器网络的最优权重估计。
This paper proposes a Taijiquan key action recognition algorithm based on perceptron network.Firstly,background subtraction is employed to detect Taijiquan player.Secondly,a multilayer perceptron network is designed to recognize Taijiquan key action.Finally,a data set of Taijiquan key action is established and the optimal weight is estimated by gradient descent method.
作者
孙瑞阳
孙玉滨
段炼
赵蓝飞
Sun Rui-yang;Sun Yu-bin;Duan Lian;Zhao Lan-fei(College of National Traditional Sports,Harbin Sport University,Heilongjiang Harbin 150008;Financial Department,Harbin Sport University,Heilongjiang Harbin 150008;The Higher Educational Key Laboratory for Measuring and Control Technology and Instrumentations of Heilongjiang Province,Harbin University of Science and Technology,Heilongjiang Harbin 150080)
出处
《中国新通信》
2021年第2期235-236,共2页
China New Telecommunications
基金
2018年黑龙江省省属本科高校基本科研业务费项目资助(2018KYYWF-PY07)
黑龙江省普通高校基本科研业务费专项资金资助(LGYC2018JC050)资助课题。
关键词
太极拳关键动作识别
感知器网络
背景差分法
Taijiquan key action recognition
perceptron network
background subtraction
