基于Spark平台的医学图像大数据研究

Research on Medical Image Large Data based on Spark Platform

笔者主要利用Spark分布式处理平台针对医学图像大数据进行研究,这对于后续进行医学图像大数据的检索或者分析处理都是非常基础并且重要的一步工作。Spark作为Hadoop生态系统中的大数据处理平台,被广泛运用于各行业大数据分析中。笔者通过参考淘宝、Skyblue等系统级别的处理方案,提出了利用Hadoop生态系统中HBase数据库来存储DICOM海量标准医学图像的设计方案。该方案除了解决了元数据节点的单点故障问题之外,也提高了用户的检索效率以及图片更新的可操作性。

The author mainly use Spark distributed processing platform to study the medical image data of the author,which is a very basic and important step for the subsequent retrieval or analysis of the medical image data.As a large data processing platform in Hadoop ecosystem,Spark has been widely used in various industries for large data analysis.By referring to the system level solutions such as Taobao and Skyblue,the author puts forward the design scheme of storing DICOM massive standard medical images by using HBase database in Hadoop ecosystem.In addition to solving the problem of single point failure of metadata nodes,this scheme also improves the retrieval efficiency and the operability of image updating.