网站首页            期刊简介             编委会             投稿指南             期刊订阅             下载中心             在线留言            联系我们             English
  2025年4月5日 星期六  
文章快速检索
中国生物医学工程学报
  论文 本期目录 | 过刊浏览 | 高级检索 |
基于极限学习机的跌倒检测分类识别研究
1 东北大学中荷生物医学与信息工程学院, 沈阳 110819
2 东北大学信息科学与工程学院, 沈阳 110819
Research on Fall Detection Classification Recognition Based on Extreme Learning Machine
1 SinoDutch Biomedical and Information Engineering School, Northeastern University, Shenyang 110819, China
2 School of Information Science and Engineering, Northeastern University, Shenyang 110819, China
全文: PDF (1558 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 由于年龄和身体条件的限制,在老年人群中跌倒是非常普遍的现象。因此,根据老年人跌倒的运动特征,远程监测他们在各个时间段的状态,以便在其摔倒或突发状况时及时采取措施显得尤为重要。针对人体运动状态进行监测,分析人体运动学特征,提出基于极限学习机的跌倒检测算法。运用三维加速度传感器采集人体的三维加速度值,建立跌倒检测特征模型。在此基础上,建立基于极限学习机的跌倒检测分类器,完成对老年人的计算机辅助跌倒检测。实验数据共540例样本,选用了不同数量的训练集和测试集,其中440例作为训练数据,其余100例为测试数据。测试结果表明,准确率为93%,敏感度为87.5%,特异性为91.7%,具有良好的分类性能。在对分类训练的运行时间方面,基于极限学习机的跌倒检测方法与传统的机器学习方法相比具有明显优势。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
王之琼1
2曲璐渲1 隋雨彤1
关键词 极限学习机 跌倒检测 BP神经网络 支持向量机    
Abstract:Falls in the elderly population is a very common phenomenon because of the limitation of age and physical condition. Therefore, it is important to remote monitor the elders state real time according to the movement characteristics of falling down, in order to take actions in time when elderly people fall down in some emergency situations. Characteristics of human body kinematics were analyzed in view of the human movement state monitoring, and an algorithm that fall detection based on extreme learning machine (ELM) was proposed. The method is to collect the body’s 3D acceleration value through a sensor. Then the model of fall detection was established. On the basis of the model, a fall detection classifier based on ELM was established and the computer aided detection for falls was completed. The 540 samples of experimental data were obtained, 440 samples for training and 100 samples for testing. The accuracy of the method was 93%, the sensitivity was 87.5%, and the specificity was 91.7%. Results show that the established method has good classification performance and advantages on the running time for classification of training in reference to the conventional machine learning.
Key wordsextreme learning machine    fall detection    BP neural network    support vector machine
    
基金资助:国家自然科学基金项目(61100022)
引用本文:   
王之琼1,2曲璐渲1    隋雨彤1   . 基于极限学习机的跌倒检测分类识别研究[J]. 中国生物医学工程学报, 2014, 33(4): 418-424.
WANG Zhi Qiong1,2QU Lu Xuan1      SUI Yu Tong1   BAO Nan1     KANG Yan1#*. Research on Fall Detection Classification Recognition Based on Extreme Learning Machine. journal1, 2014, 33(4): 418-424.
链接本文:  
http://cjbme.csbme.org/CN/10.3969/j.issn.0258-8021. 2014.04.05     或     http://cjbme.csbme.org/CN/Y2014/V33/I4/418
版权所有 © 2015 《中国生物医学工程学报》编辑部
本系统由北京玛格泰克科技发展有限公司设计开发