融合机器视觉与惯性传感器的智能康复评估技术

Intelligent rehabilitation assessment technology integrating machine vision and inertial measurement sensors

为缓解患有运动功能障碍的患者不方便频繁前往医院进行康复评估的问题,本文提出一种基于“互联网+”的居家康复评估技术,为康复医师提供参考,节约康复医师和患者的时间,缓解医疗资源不足。针对脑卒中患者上肢康复量表的评估问题,仅使用惯性传感器或者机器视觉建立的长短期记忆人工神经网络分类模型在测试样本中准确率分别为55.6%、92.6%,本文在基于机器视觉方式获取人体的3D坐标的同时,通过惯性传感器获取肢体的方向并融合两者的数据,进一步采用长短期记忆人工神经网络分类模型对数据进行分析,在测试样本中取得了98.1%的分类准确率,取得了良好的实验效果。

In order to alleviate the problem that patients with motor dysfunction are inconvenient to frequently go to the hospital for rehabilitation evaluation,this paper proposes a home rehabilitation evaluation technology based on"Internet+",which provides reference for rehabilitation physicians,saves the time of rehabilitation physicians and patients,and alleviates Insufficient medical resources.For the evaluation of the upper limb rehabilitation scale for stroke patients,the long-term and short-term memory artificial neural network classification model established only by inertial measurement sensors or machine vision has an accuracy rate of 55.6%and 92.6%in the test samples,respectively.At the same time as the 3 D coordinates of the human body,the orientation of the limb is obtained through the inertial sensor and the data of the two are fused.The long short-term memory artificial neural network classification model is further used to analyze the data,and the classification accuracy rate of 98.1%in the test sample is obtained.good experimental results.