摘要
针对海量医学图像存储及检索效率低的问题,该文提出利用Hadoop平台分别实现大量医学图像的分布式存储以及并行处理模式下的基于内容的医学图像检索。利用HIPI图像处理接口将医学图像上传到分布式文件系统(HDFS)中;然后,分别提取图像的形状以及纹理特征,并将其特征向量存储到HDFS中;最后,利用Map Reduce模型实现并行式检索并将图像检索结果按照相似度大小进行排序及显示。实验结果表明,在Hadoop云平台下大量医学图像的存储效率以及检索效率较高,且图像数量越多效率优势越明显。
Since the storage and retrieval efficiency of massive medical images is low,the Hadoop platform is used to realize the distributed storage of the massive medical images and content?based medical image retrieve in the parallel processing mode respectively. The Hadoop image processing interface(HIPI)is adopted to upload the medical images to the Hadoop distributed file system(HDFS). The shape and texture features of the images are extracted respectively,and their feature vectors are stored in HDFS. The MapReduce model is employed to realize the parallel retrieval,and sort and display the image retrieval results ac?cording to their similarity. The experimental results show that the efficiency of the massive medical images storage and retrieval by means of Hadoop cloud platform is high,and the efficiency advantage is significantly obvious with the increase of the image quantity.
作者
郝娟
吕晓琪
温秀梅
谷宇
黄显武
HAO Juan;Lü Xiaoqi;WEN Xiumei;GU Yu;HUANG Xianwu(School of Information Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,China;School of Information Engineering,Hebei University of Architecture,Zhangjiakou 075000,China)
出处
《现代电子技术》
北大核心
2017年第4期115-119,共5页
Modern Electronics Technique
基金
国家自然科学基金资助项目(61179019
61261028)
内蒙古自治区自然科学基金项目(2014MS0828
2015MS0604)
内蒙古科技大学创新基金(2014QDL045)
内蒙古自治区高等学校科学研究项目(NJZY145)
河北建筑工程学院科研基金项目(ZD201411)
作者简介
郝娟（1989—），女，河北保定人，硕士。研究方向为医学图像处理;吕晓琪（1963—），男，内蒙古包头人，教授，博士生导师。主要研究方向为医学图像处理。
