《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (2): 622-630.DOI: 10.11772/j.issn.1001-9081.2021040618

• 前沿与综合应用 • 上一篇    

基于关节点特征的跌倒检测算法

曹建荣1,2, 朱亚琴1(), 张玉婷1, 吕俊杰1, 杨红娟1,2   

  1. 1.山东建筑大学 信息与电气工程学院,济南 250101
    2.山东省智能建筑技术重点实验室(山东建筑大学),济南 250101
  • 收稿日期:2021-04-20 修回日期:2021-07-16 接受日期:2021-07-23 发布日期:2022-02-21 出版日期:2022-02-10
  • 通讯作者: 朱亚琴
  • 作者简介:曹建荣(1965—),男,山东济南人,教授,博士,主要研究方向:模式识别与智能信息处理、视频分析、深度学习、数据挖掘;
    朱亚琴(1995—),女,山东济南人,硕士研究生,主要研究方向:模式识别与智能信息处理、视频分析、深度学习;
    张玉婷(1997—),女,山东济南人,硕士研究生,主要研究方向:深度学习、视频总结、图像处理;
    吕俊杰(1991—),男,山东济南人,硕士研究生,主要研究方向:模式识别与智能信息处理、视频分析、深度学习、数据挖掘;
    杨红娟(1978—),女,山东济南人,副教授,博士,主要研究方向:模式识别与智能信息处理、深度学习、数据挖掘。
  • 基金资助:
    山东省重点研发计划项目(2019GSF111054);山东省重大科技创新工程(2019JZZY010120)

Fall detection algorithm based on joint point features

Jianrong CAO1,2, Yaqin ZHU1(), Yuting ZHANG1, Junjie LYU1, Hongjuan YANG1,2   

  1. 1.School of Information and Electrical Engineering,Shandong Jianzhu University,Jinan Shandong 250101,China
    2.Shandong Provincial Key Laboratory of Intelligent Buildings Technology (Shandong Jianzhu University),Jinan Shandong 250101,China
  • Received:2021-04-20 Revised:2021-07-16 Accepted:2021-07-23 Online:2022-02-21 Published:2022-02-10
  • Contact: Yaqin ZHU
  • About author:CAO Jianrong, born in 1965, Ph. D., professor. His research interests include pattern recognition intelligent information processing, video analysis, deep learning, data mining.
    ZHU Yaqin, born in 1995, M. S. candidate. Her research interests include pattern recognition and intelligent information processing, video analysis, deep learning.
    ZHANG Yuting, born in 1997, M. S. candidate. Her research interests include deep learning, video summary, image processing.
    LYU Junjie, born in 1991, M. S. candidate. His research interests include pattern recognition and intelligent information processing, video analysis, deep learning, data mining.
    YANG Hongjuan, born in 1978, Ph. D., associate professor. Her research interests include pattern recognition and intelligent information processing, deep learning, data mining.
  • Supported by:
    Shandong Provincial Key Research and Development Program(2019GSF111054);Shandong Province Major Scientific and Technological Innovation Project(2019JZZY010120)

摘要:

针对跌倒检测算法中存在网络计算量大和类跌倒行为难以区分的问题,提出一种基于关节点特征的跌倒检测算法。首先,在目前先进的CenterNet算法基础上提出了深度可分离卷积CenterNet (DSC-CenterNet)关节点检测算法,从而在减少骨干网络计算量的同时准确检测人体关节点并获取关节点坐标;然后,基于关节点位置和人体先验知识来提取可充分表达跌倒行为的空间特征和时间特征作为关节点特征;最后,把关节点特征向量输入全连接层,并经Sigmoid分类器输出跌倒或非跌倒两种类别,从而实现人体目标的跌倒检测。实验结果表明,所提算法在UR Fall Detection数据集上对不同状态变化下跌倒检测的平均准确率达到98.00%,区分类跌倒行为的准确率达到98.22%,跌倒检测速度为18.6 frame/s。与原CenterNet结合关节点特征跌倒检测的算法相比,DSC-CenterNet结合关节点特征算法的跌倒检测速度提升了22.37%,提高后的速度可有效满足视频监控下人体跌倒检测任务的实时性。该算法能有效提高跌倒检测速度并对人体跌倒状态进行准确检测,且进一步验证了基于关节点特征的跌倒检测算法在视频跌倒行为分析中的可行性与高效性。

关键词: 跌倒检测, 深度学习, CenterNet算法, 关节点检测, 关节点特征

Abstract:

In order to solve the problems of large amount of network computation and difficulty in distinguishing falling-like behaviors in fall detection algorithms, a fall detection algorithm based on joint point features was proposed. Firstly, based on the current advanced CenterNet algorithm, a Depthwise Separable Convolution-CenterNet (DSC-CenterNet) joint point detection algorithm was proposed to accurately detect human joint points and obtain joint point coordinates while reducing the amount of backbone network computation. Then, based on the joint point coordinates and prior knowledge of the human body, the spatial and temporal features expressing the fall behavior were extracted as the joint point features. Finally, the joint point feature vector was input into the fully connected layer and processed by Sigmoid classifier to output two categories: fall or non-fall, thereby achieving the fall detection of human targets. Experimental results on UR Fall Detection dataset show that the proposed algorithm has the average accuracy of fall detection under different state changes reached 98.00%, the accuracy of distinguishing falling-like behaviors reached 98.22% and the fall detection speed of 18.6 frame/s. Compared with the algorithm of the original CenterNet combining with joint point features, the algorithm of DSC-CenterNet combining with joint point features has the average detection accuracy increased by 22.37%. The improved speed can effectively meet the realtime requirement of the human fall detection tasks under surveillance video. This algorithm can effectively increase fall detection speed and accurately detect the fall state of human body, which further verifies the feasibility and efficiency of fall detection algorithm based on joint point features in the video fall behavior analysis.

Key words: fall detection, deep learning, CenterNet algorithm, joint point detection, joint point feature

中图分类号: