Massive medical image retrieval system based on Hadoop
FAN Min1,XU Shengcai2
1. Department of Informatics and Electronics, Hangzhou Vocational and Technical College, Hangzhou Zhejiang 310018, China 2. Department of Electronics and Information Engineering, Tongji University, Shanghai 201815, China
Abstract:In order to improve the retrieval efficiency of massive medical images, a new medical image retrieval system was proposed based on distributed Hadoop to solve the low efficiency of medical image retrieval system based on single node. Firstly, the features of medical image were extracted by using Brushlet transform and Local Binary Pattern (LBP) algorithm, and the feature database was stored in the Hadoop Distributed File System (HDFS). Secondly, the Map was used to match the features of retrieval images and medical images in the library, and the matching results of the Map task were collected and sorted by the Reduce function. Finally, the optimum results of medical image retrieval were obtained according to the ordering. The test results show that, compared with other medical image retrieval systems, the proposed system reduces the time of image storage and retrieval, and improves the image retrieval speed.
范敏 徐胜才. 基于Hadoop的海量医学图像检索系统[J]. 计算机应用, 2013, 33(12): 3345-3349.
FAN Min XU Shengcai. Massive medical image retrieval system based on Hadoop. Journal of Computer Applications, 2013, 33(12): 3345-3349.