%0 Journal Article
%T 基于OpenPose的脑卒中肢体康复训练评估辅助系统开发<br>Development of an OpenPose-Based Assessment Aid System for Stroke Physical Rehabilitation Training
%A 刘肇川
%A 毛文婷
%A 陈炫
%J Artificial Intelligence and Robotics Research
%P 299-307
%@ 2326-3423
%D 2022
%I Hans Publishing
%R 10.12677/AIRR.2022.113031
%X 针对脑卒中患者的居家肢体康复训练缺乏指导、看护者的康复知识欠缺的问题，开发一套针对脑卒中患者肢体康复训练的智能化辅助系统。通过摄像头获取患者康复训练过程中的图像，利用OpenPose与MoveNet姿态检测库与提取患者训练过程的实时动作骨架坐标序列，构建用户肢体运动评价模型，对康复训练者的训练过程进行评价，得出训练者的康复训练质量，并通过文字提示的方式提醒用户不达标的动作以及改正方式，指导和引导患者逐步达到康复训练的标准。<br />
To address the problems of lack of guidance and caregiver’s knowledge in home physical rehabilitation training for stroke patients, an intelligent assisted system for physical rehabilitation training of stroke patients was developed. The system uses the OpenPose and MoveNet posture detection libraries to extract the real-time skeleton coordinate sequences of the patient’s training process, construct a user’s limb movement evaluation model, evaluate the training process of the trainer, and derive the quality of the trainer’s rehabilitation training, and remind the user of the substandard movements and the correction methods through text prompts. It also reminds the user of the substandard movements and the ways to correct them by means of textual prompts, and guides the patient to gradually achieve the rehabilitation training standard.
%K 计算机视觉，神经网络，OpenPose算法，MoveNet算法<br>Computer Vision
%K Neural Networks
%K OpenPose Algorithm
%K MoveNet Algorithm
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=54801