基于神经网络的动态手势识别方法研究  

Research on the Dynamic Gesture Recognition Method Based on the Neural Network

作　　者：秦利娟 王礼云 QIN Lijuan;WANG Liyun(Zhengzhou University of Industry Technology,Zhengzhou Henan 450000)

机构地区：[1]郑州工业应用技术学院,河南郑州450000

出　　处：《软件》2024年第9期175-177,共3页Software

基　　金：程序设计基础课程思政教学团队(郑工政[2022]6号);2022年度郑州地方高校急(特)需专业建设项目(郑教高函[2023]42号);基于云仿真平台simdroid的“双师型”师资培训(220906517024221);郑州工业应用技术学院课程思政样板课-Python语言基础。

摘　　要：人工神经网络及计算机视觉分析技术的发展使得动态手势识别技术越来越成熟,被广泛应用于体育运动训练、机器人设计等多个领域。传统的手势识别模型需要人工提取特征,无法实现复杂特征的人工操作,导致手势识别效果不佳。深度学习在目标识别、分类任务等领域的应用越来越广泛,神经网络技术具备强大的图像特征提取能力,可获得更好的手势识别效果。本研究提出一种基于卷积神经网络的动态手势识别模型。The development of artificial neural network and computer vision analysis technology makes the dynamic gesture recognition technology more and more mature,which is widely used in sports training,robot design and other fields.The traditional gesture recognition model requires manual extraction of features,which fails to realize the manual operation of complex features,resulting in poor gesture recognition effect.Deep learning is more and more widely used in the fields of target recognition and classification tasks.Neural network technology has powerful image feature extraction ability and can obtain better gesture recognition effect.This study proposes a dynamic gesture recognition model based on the convolutional neural network.

关 键 词：卷积神经网络 动态手势 图像识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]