摘要
随着中国老龄化社会的到来,应对老年人口安全问题,特别是摔倒问题,变得越来越重要。提出了一种基于卷积神经网络(CNN)和多特征融合的预测系统。该系统整合了图像和生理信号等多种类型的特征信息,以提高摔倒预测的准确性。实验验证了基于CNN的多模型结构在老人摔倒预测中的优越性,以及多特征融合策略对模型性能的提升作用。与其他方法相比,所提出的方法在准确率、召回率、精确率和F1分数方面表现出优越性,准确率可达到95.93%。此研究为预测和预防老年人摔倒提供了一种高效且可靠的方法。
As China enters an aging society,addressing safety issues for the elderly population,especially the problem of falling has become increasingly important.A prediction system based on Convolutional Neural Networks(CNN)and multi-feature fusion has been proposed.This system integrates various types of feature information,such as images and physiological signals,to improve the accuracy of falling prediction.Experiments have validated the superiority of the multi-model structure based on CNN in predicting the elderly man falling and the enhancement of model performance by the multi-feature fusion strategy.Compared to other methods,the proposed method demonstrates superior performance in terms of accuracy,recall,precision,and F1 score,with an accuracy of 95.93%.This research provides an efficient and reliable method for predicting and preventing falling among the elderly man.
作者
胡昕
刘瑞安
黄玉兰
任超
徐宇辉
HU Xin;LIU Rui-an;HUANG Yu-lan;REN Chao;XU Yu-hui(College of Electronical and Information Engineering,Tianjin Normal University,Tianjin 300380,China)
出处
《信息技术》
2024年第10期94-101,共8页
Information Technology
基金
天津师范大学研究生科研创新项目资助(2022KYCX-105Y)。
关键词
CNN算法
多特征融合
特征提取
老人摔倒预测
数据集
CNN algorithm
multi-feature fusion
feature extraction
prediction of elderly man falling
data set
