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.
关键词
时空图卷积网络(ST-GCN)/儿童坐姿识别/学习桌椅
Key words
Spatio-Temporal Graph Convolutional Network(ST-GCN)/Children's sitting posture recognition/earning desk and chair
维普
万方数据