基于hadoop技术的医疗云数据安全管理系统设计与应用  

Design And Application of Medical Cloud Data Security Management System Based on Hadoop Technology

作　　者：赵沛[1] 曹郡[1] 贾瑞龙 Zhao Pei;Cao Jun;Jia Rui-long(JiangSu Province Hospitalf The First Affiliated Hospital With Nanjing Medical University,2100291 China;People's Hospital of Kizilsu Kirgiz Autonomous Prefecture,Kizilsu Kirgiz 845350,China)

机构地区：[1]江苏省人民医院(南京医科大学第一附属医院),江苏南京210029 [2]克孜勒苏柯尔克孜自治州人民医院,新疆克孜勒苏柯尔克孜845350

出　　处：《生命科学仪器》2020年第3期77-82,65,共7页Life Science Instruments

基　　金：新疆克州3123人才课题,编号:RK2019094。

摘　　要：医疗云数据管理系统由文本、数字、图片和视频等不同格式的组成,数据量大且无规定结构。为了保证数据安全,设计了一种基于云技术的医疗数据管理系统,在hadoop MapReduce框架上,通过生物认证技术,使用弹性反向传播神经网络执行用于身份验证的签名样本训练。实验结果标明,与现有的其他方法相比,该方法可实现加速效果提升,误差率(EER)为0.12,灵敏度为0.98,特异性为0.95。该方法是医疗数据管理的有效解决方案,为医疗云数据安全管理技术提供了新思路。The medical cloud data management system is composed of text,digital,picture and video formats,with large amount of data and no specified structure.In order to ensure data security,a medical data management system based on cloud technology is designed.In the Hadoop MapReduce framework,through the biometric authentication technology,the signature sample training for authentication is performed using the elastic back propagation neural network.The experimental results show that compared with other methods,this method can achieve acceleration,the error rate(EER)is 0.12,the sensitivity is 0.98,and the specificity is 0.95.This method is an effective solution for medical data management and provides a new idea for medical cloud data security management technology.

关 键 词：生物识别 身份验证 医疗云 HADOOP 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]