基于图卷积网络的儿童坐姿检测学习桌椅设计方法研究

Research on the Design Method of Children's Sitting Posture Detection Study Table and Chair Based on Graph Convolutional Network

儿童长期处于伏案学习的状态,不良坐姿对儿童生长发育容易造成严重影响,对儿童坐姿的矫正已刻不容缓。通过对儿童在学习桌椅上的坐姿行为调研,基于OpenPose姿态估计算法进行坐姿检测分析,引入ST-GCN模型进行坐姿识别与评判,结果表明:使用ST-GCN模型能够快速准确识别儿童的八种坐姿,并根据识别结果对儿童进行有效的错误坐姿提示,其Macro-F1和Micro-F1评价指标分别提高了6.8%和7.4%。同时表明儿童坐姿矫正在自适应儿童学习桌椅上应用的可行性及有效性,可为智能儿童课桌椅的设计提供技术支撑。

Children are in the state of studying for a long time,and the bad sitting posture has a serious impact on children's growth and development,and the correction of children's sitting posture is urgent.Through the research of children's sitting behavior on the study table and chair,based on the OpenPose posture estimation algorithm for sitting posture detection and analysis,the introduction of the ST-GCN model for sitting posture recognition and evaluation,the results show that:the introduction of the ST-GCN model can quickly and accurately identify the children's eight sitting postures,and according to the recognition results of the children's effective tips for the incorrect sitting postures,and the Macro-F1 and Micro-F1 evaluation indexes are improved by 6.8%and 7.4%,respectively.Micro-F1 evaluation indexes were improved by 6.8%and 7.4%respectively.At the same time,it shows the feasibility and effectiveness of children's sitting posture correction applied on adaptive children's learning chair,which provides algorithm and function basis for the design of intelligent children's sitting posture correction furniture.