作　　者： (郭文亮）;

机构地区： 邯郸市中心医院信息科,河北邯郸056001

出　　处： 《中国医疗设备》 2017年第9期115-117,129,共4页

摘　　要： 目的医学影像在临床诊断、科学教研中发挥着越来越重要的作用,本文探讨通过云平台解决医学影像数据量剧增、检索效率低等问题的可行性。方法将传统的光纤存储局域网络集中式存储和分布式文件系统(Hadoop Distributed File System,HDFS)存储相结合,形成一种混合式存储架构。其中通过文件序列化的方式解决HDFS不适合医学影像小文件存储的问题,通过集中式存储解决医学影像实时性访问的问题。结果随着客户端数量增加,在混合式存储方式下,医学影像数据的读、写速度较传统的集中式存储都有明显的提高。结论采用Hadoop平台的混合式存储方式可以满足剧增的医学影像数据的存储要求,并且可以提高数据的读写速度。 Objective Medical image is playing a more and more important role in clinical diagnosis and scientifc research. This paper aimed to investigate the feasibility of solving the problem of soaring medical imaging data and ineffective searching by cloud platform. Methods This paper proposed a hybrid storage architecture which combined the traditional fibro optical storage area network centralized storage with hadoop distributed fle system （HDFS） distributed storage. Through the fle serialization solution, this paper solved the problem that HDFS storage was not suitable for the small medical image. Through centralized storage the paper solved the problem of real-time access to medical image. Results With the increasing of client, reading and writing speed of medical image data in the mixed storage mode was obviously improved compared with the traditional centralized storage. Conclusion The hybrid storage mode using the Hadoop platform can meet requirement of the roaring medical image data storage, and improve the reading and writing speed of data.

关 键 词： 医学影像 小文件存储 实时性访问 医学数字成像和通信标准