摘要
鉴于青少年在线学习的不断普及,本文提出使用计算机深度学习技术进行在线学习时特定动作识别研究的方法。该方法应用卷积神经网络,基于帧间差分法提取视频关键帧,融合了空间和时间特征,用于识别青少年在线学习时不同类型的动作。经实验证实,该方法能有效识别出躺卧、站立、移动等特定动作,验证了该方法的有效性。
In view of the growing popularity of online learning among teenagers,This paper presents a research method about specific motion recognition method based on learning of youth online using computer in-depth learning technology.In this method,convolution neural network combing spatial and time features is used to identify different types of actions of teenagers in online learning extracting video key frames by the frame difference method.Experiments show that the method can effectively recognize the specific actions such as lying,standing and moving,and the effectiveness of the method was verified.
作者
徐丽珍
吴功才
XU Lizhen;WU Gongcai(Hangzhou Vocational&Technical College,Hangzhou Zhejiang 310006,China)
出处
《信息与电脑》
2021年第13期24-26,共3页
Information & Computer
基金
2020年度国内访问工程师校企合作项目“基于卷积神经网络的视频动作识别研究”(项目编号:FG2020087)
2020年度浙江省教育厅科研课题“基于OpenCV的儿童视力保护系统设计”(项目编号:Y202045274)。
关键词
空间特征
时间特征
动作识别
spatial features
time features
motion recognition