基于计算机视觉的儿童运动发展辅助诊断技术研究  

Research on computer vision-based auxiliary diagnosis technique for children's motor development

作　　者：舒大伟 雷毅 陈艳杰 王青[4] 臧林 杨吉江[4] SHU Dawei;LEI Yi;CHEN Yanjie;WANG Qing;ZANG Lin;YANG Jijiang(Tsinghua Shenzhen International Graduate School,Shenzhen 518055,Guangdong Province,China;School of Software Engineering,Faculty of Information Technology,Beijing University of Technology;Children's Health Center,Beijing Children's Hospital Affiliated to Capital Medical University;Department of Automation,Tsinghua University;Cross-strait Tsinghua Research Institute)

机构地区：[1]清华大学深圳国际研究生院,广东深圳518055 [2]北京工业大学信息学部软件学院 [3]首都医科大学附属北京儿童医院儿童保健中心 [4]清华大学自动化系 [5]清华海峡研究院

出　　处：《中国数字医学》2023年第1期43-47,共5页China Digital Medicine

基　　金：国家重点研发计划(2020YFC2006702);北京市朝阳区协同创新项目(CYXC2010)。

摘　　要：目的:提出一套可客观量化评估儿童运动发展的辅助诊断方法,对儿童运动发展状况进行定量评估,辅助诊断儿童发育性协调障碍。方法:以规定动作类别的儿童动作视频数据作为输入,使用关键点检测算法得到人体关键点数据,在进行动作规整后使用基于相似度计算方法得到对应动作的质量评估结果,通过动作完成质量反映发育性协调障碍诊断结果。结果:结合关键点波形分析以及动作质量评估,得出的结果与观察得到的结果表现出明显的一致性。结论:本研究以动作视频为基础结合计算机视觉方法将儿童发育性协调障碍的诊断方法进行标准化,为该疾病的诊断以及后续干预效果评估提供了一种标准量度。Objective In order to carry out quantitative assessment and research on children's motor development,and to diagnose the developmental coordination disorder(DCD),a set of auxiliary diagnostic methods were proposed which can objectively and quantitatively assess children's motor development.Methods With the child motion video data of the specified categories as the input,the key point detection algorithm was used to obtain the key point data of the human body.After the motions were regularized,the quality assessment results of the corresponding motions were obtained by using the template similarity calculation method,and the diagnosis results of DCD were reflected by the motions'completion quality.Results Based on the key points of waveform analysis and motion quality evalution,the observed and obtained results showed clear consistency.Conclusion Based on motion videos using computer vision methods,this paper standardizes the DCD diagnosis methods and provides a standard measure for the DCD diagnosis and assessment of subsequent intervention effects.

关 键 词：发育性协调障碍 人体关键点检测 动作质量评估 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]