收稿日期:2019-10-21 修回日期:2020-02-24
www yingyuxuexi net
基金项目：教育部科技发展中心产学研创新基金(2018A01002)；教育部人文社科基金项目(15YJC890004)；中国博士后基金项目(2017M610852)；吉林省社科基金重点项目(2016A5)作者简介：李晓峰(1978-)，男，博士研究生，教授,CCF 高级会员(16287S)，研究方向为数据挖掘㊁人工智能㊁智能交通㊁社会计算㊁智慧医疗㊁
体育工程学㊂基于卷积神经网络的运动视频可靠性评估算法
yue是什么意思hole是什么意思李晓峰1，邢金明2
(1.黑龙江外国语学院信息工程系，黑龙江哈尔滨150025;2.东北师范大学，吉林长春130024)摘 要：为了提高对人体运动视频的自动识别和检测能力，提高可靠性能，提出一种基于卷积神经网络的人体运动视频传输可靠性评估算法㊂由卷积神经网络算法识别人体运动视频传输的自适应分类，提取人体运动视频的空间边缘像素点分布标，结合边缘模块特征匹配技术构建人体运动视频的分块检测模型，实现对人体运动视频的特征辨识和图像采样㊂采用Harris 角点检测方法定位人体运动视频的分块区域，在人体运动视频的分块区域内检测人体运动视频的形体轮廓区域，构建可靠性评估均衡博弈模型完成视频干扰抑制㊂采用视频特征提取和自动降噪方法分离人体运动视频传输过程中的多径特征，在神经网络的隐含层引入人体运动视频的几何特征，得到人体运动视频传输的可靠性评估的学习系数，完成可靠性评估㊂实
验结果表明，采用该方法进行人体运动视频传输的可靠性较好，对视频图像的特征分辨能力较强且视频图像传输耗时较短，降低了视频传输的误码率㊂
汉语翻译英语转换器关键词：卷积神经网络；人体运动；视频传输；可靠性评估；误码率
中图分类号:TP391 文献标识码:A 文章编号:1673-629X (2020)09-0071-06
doi:10.3969/j.issn.1673-629X.2020.09.013
Reliability Evaluation Algorithm of Motion Video Bad on Convolution Neural Network
LI Xiao -feng 1,XING Jin -ming 2
(1.Department of Information Engineering ,Heilongjiang International University ,Harbin 150025,China ;2.Northeast Normal University ,Changchun 130024,China )Abstract :In order to improve the ability of automatic recognition and detection of human motion video ,as well as the reliability ,we propo a reliability evaluation algorithm of human motion video transmission bad on convolution neural network.The adaptive classi⁃fication of human motion video transmission is recognized by convolution neural network algorithm ,and the spatial edge pixel distribution mark of human motion video is extracted.Combined with the edge module feature matching technology ,the
block detection model of human motion video is constructed ,and the feature identification and image sampling of human motion video are realized.The Harris corner detection method is ud to locate the gmented area of human motion video ,and the shape outline area of human motion video is detected in the gmented area of human motion video.The reliability evaluation equilibrium game model is constructed to complete the video interference suppression.The multi -path features in the process of human motion video transmission are parated by video feature extraction and automatic noi reduction.The geometric features of human motion video are introduced into the hidden layer of neural network ,and the learning coefficient of reliability evaluation of human motion video transmission is obtained ,and the reliability evaluation is completed.The experiment shows that the propod method is reliable ,the feature resolution of video image is strong and the transmission time of video image is short ,which reduces the bit error rate of video transmission.
Key words :convolution neural network ;human motion ;video transmission ;reliability evaluation ;bit error rate
新目标英语八年级上册0 引 言
人体运动属于复杂且普遍的现象，日常生活中的跑步㊁走路以及机体呼吸，都属于人体运动的一部分
none of that商店的英文㊂人体运动视频的巨大数据量成为其存储与传输的瓶第30卷 第9期2020年9月 计算机技术与发展COMPUTER TECHNOLOGY AND DEVELOPMENT矮个子衣服的穿配法
Vol.30 No.9Sep. 2020
萱萱 bleeding love>gyz
