基于Stacked Hourglass的智能康复评估系统研究  

Research on intelligent rehabilitation evaluation system based on Stacked Hourglass

作　　者：卢鹏飞 郑建立[1] 尹梓名[1] LU Pengfei;ZHENG Jianli;YIN Ziming(School of Medical Instrument and Food Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学医疗器械与食品学院,上海200093

出　　处：《智能计算机与应用》2021年第2期60-63,68,共5页Intelligent Computer and Applications

基　　金：国家重点研发计划(2019YFC2005801,2019YFC2005800)。

摘　　要：2020年康复领域异常严峻,一方面,疫情爆发,人群不易聚集,患者不方便去医院或者康复中心做康复及检测身体状况,另一方面,全球人口老龄化问题加剧,老人腿脚不方便,不便于频繁地去康复中心复查,并且全国康复中心资源较少,不仅造成康复师工作压力增大,还会出现医患沟通不充分,延误康复等问题。随着人工智能技术不断发展,机器视觉在人工智能中的表现尤为突出,其中骨架识别不断被认可,应用在了多个领域。因此本次研究设计了基于Stacked Hourglass的智能康复评估系统,期望可以提高康复医师工作效率。本系统主要为智能康复评估,首先构建以量表为中心,建立量表答题结构的数据库,将量表储存在数据库中。其次,将骨架识别移植到移动端软件识别出量表所需数据。然后,将骨架识别数据通过计算,算出填表所需值,智能填表。最后,通过各个量表数据,全部显示出来并提供给康复师,让康复师了解患者状况。通过这些步骤设计量表模型,完成智能康复评估。The field of rehabilitation in 2020 is extremely severe.On one hand,the outbreak of the epidemic makes it difficult for people to gather,and it is inconvenient for patients to go to hospitals or rehabilitation centers for rehabilitation and check their physical conditions.On the other hand,the problem of global population aging has intensified,and the legs and feet of the elderly are inconvenient.Not only frequent visits to rehabilitation centers for review and fewer resources in rehabilitation centers across the country increase the work pressure of rehabilitation practitioners,but also inadequate communication between doctors and patients and delays in rehabilitation will also happen.With the continuous development of artificial intelligence technology,the performance of machine vision in artificial intelligence is particularly prominent,among which skeleton recognition is continuously recognized and applied in many fields.Therefore,this study designs an intelligent rehabilitation evaluation system based on Stacked Hourglass,hoping to improve the efficiency of rehabilitation physicians.This system is mainly for intelligent rehabilitation assessment.First,build a database with the scale as the center,establish the structure of the scale,and store the scale in the database.Secondly,transplant the skeleton recognition to the mobile terminal software to recognize the data required by the scale.Then,the skeleton identification data is calculated to achieve the required value for filling the form,and the form is filled intelligently.Finally,through the data of each scale,all are displayed to the rehabilitation practitioner,so that the rehabilitation practitioner can understand the patient’s condition.Through these steps,the scale model is designed to complete the intelligent rehabilitation assessment.

关 键 词：骨架识别 HOURGLASS TensorFlow 智能量表 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]