机器学习在医疗设备故障识别与预测中的应用与展望  

Application and prospect of machine learning in identification and prediction of medical equipment

作　　者：陈晓宇 王子洪 郭海涛 黄小龙 陈文勤 Chen Xiaoyu;Wang Zihong;Guo Haitao;Huang Xiaolong;Chen Wenqin(Department of Medical Engineering,The First Hospital Affiliated to Army Medical University Chongqing 400038,China;Office of Medicine,The First Hospital Affiliated to Army Medical University Chongqing 400038,China;Laboratory,Shuangqiao Economic and Technological Development Zone People's Hospital,Chongqing 400900,China)

机构地区：[1]陆军军医大学第一附属医院医学工程科,重庆400038 [2]陆军军医大学第一附属医院医疗办公室,重庆400038 [3]重庆市双桥经济技术开发区人民医院检验科,重庆400900

出　　处：《中国医学装备》2025年第1期143-149,共7页China Medical Equipment

基　　金：重庆市科卫联合医学科研面上项目(2022MSXM060);重庆市技术创新与应用发展专项面上项目(cstc2019jscxmsxmX0183)。

摘　　要：传统的医疗设备故障识别与预测主要依靠设备管理者经验知识,无法量化且效率较低,医疗设备故障预测不准确的问题尤为突出。随着计算机技术以及机器学习技术的发展,以往需凭借经验进行的故障识别与预测,可借助机器学习方法处理故障特征以提高效率,有望弥补医疗设备故障识别与预测的学科空白。总结国内外机器学习在医疗设备和类似医疗设备的其他电子设备故障识别与故障预测中的应用情况,并根据故障识别和预测的关键技术提出相应的设计架构建议,根据医疗设备故障的特征归纳总结各机器学习的算法在识别与预测中的场景、准确率信息等信息。The conventional identification and prediction for failures of medical equipment mainly depend on experience and knowledge of manager for equipment,which are not able to be quantified and have lower efficiency.Therefore,it is obvious that the prediction for the failure of medical equipment is not accurate.With the technical development of computer and machine learning,the conventional identification and prediction that depend on experiences can deal with characteristics of failures through machine learning method to improve efficiency,which are hopeful in filling the gap of discipline about the identification and prediction for failures of medical equipment.This article summarized the application situation of machine learning in identifying and predicting failures of the medical equipment and the similarly electric equipment at home and abroad.Based on the key technique of identification and prediction,this article proposed suggestion about corresponding design architecture.According to the characteristics of the failure of medical equipment,this article summarized a series of information about algorithms of various machine learning in scene and accurate rate of identification and prediction,so as to provide references for relevant research of this field.

关 键 词：医疗设备 故障识别 故障预测 机器学习 神经网络 深度学习 

分 类 号：R197.39[医药卫生—卫生事业管理]