基于计算机视觉的儿童精细动作评估研究  

A Study of Fine Motor Assessment in Children Based on Computer Vision

作　　者：崔文波 宋文爱[1] 雷毅 王青[3] 陈艳杰 CUI Wenbo;SONG Wenai;LEI Yi;WANG Qing;CHEN Yanjie(North University of China,Taiyuan 030051,Shanxi,China)

机构地区：[1]中北大学软件学院,山西省太原市030051 [2]北京工业大学信息学部,北京市100124 [3]清华大学自动化系,北京市100084 [4]国家儿童医学中心(首都医科大学附属北京儿童医院保健中心),北京市100045

出　　处：《中国卫生信息管理杂志》2023年第4期525-531,共7页Chinese Journal of Health Informatics and Management

基　　金：慢性病防治与健康教育科研项目“基于AI技术的儿童动作发展范式研究”(项目编号:BJMB0012022028020)。

摘　　要：目的 提出一种基于计算机视觉的方法来评估儿童精细动作,以促进对儿童发育性协调障碍的准确诊断和进一步研究。方法 利用计算机视觉技术从视频中提取人体关键点,并对其进行时间序列特征提取,然后使用深度学习神经网络对儿童精细动作进行分类和预测。结果 实验结果显示,本研究方法的准确率达到81%,相较于传统的基于手工特征的评价方法,在准确率上有所提高。结论 基于计算机视觉的儿童精细动作自动评估方法,为儿童发育性协调障碍疾病的辅助诊断提供了一个有效的判别工具,为促进幼儿早期综合发展提供了有力的技术支持。Objective The aim of this paper is to propose a computer vision-based method to evaluate fine motor in children to facilitate the accurate diagnosis and further study of developmental coordination disorders in children.Methods The key points of the human body are extracted from the video using computer vision technology,and the time series features are extracted,and then the deep learning neural network is used to classify and predict the fine movements of children.Results By comparing the experimental results,the accuracy of the method in this study is up to 81%,which is higher than that of the traditional evaluation method based on manual features.Conclusion The automatic fine motor assessment method based on computer vision provides an effective discrimination tool for the auxiliary diagnosis of children with developmental coordination disorder,and provides good technical support for promoting early comprehensive development of children.

关 键 词：儿童发育性协调障碍 深度学习 精细动作 人体姿态估计 

分 类 号：R-34[医药卫生] R319