...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Philippine Information Technology >A Medical Image Retrieval Application Using Grid Technologies To Speed Up Feature Extraction
【24h】

A Medical Image Retrieval Application Using Grid Technologies To Speed Up Feature Extraction

机译：利用网格技术加快特征提取的医学图像检索应用

获取原文
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Medical image data is produced in ever–increasing quantities and varieties. The digital production of these data makes them accessible for further automatic analysis and processing. Whereas automatic analysis is fairly common in the text domain, analysis of medical images in large quantities and in a large variety is a relatively new discipline. Computational limits are often restricting the possibilities to analyze the large amounts of data produced automatically. Grid computing opens new possibilities to use an intra–hospital computing infrastructure for research projects. This article describes the griddification of a content–based image retrieval system called the GNU Image Finding Tool (GIFT). The goal of this study was to show the potential of grid computing and the benefits for the medical applications from the available grid computing power. We use the ARC (Advance Resource Connector) middleware to benefit from the distributed computing power available through the KnowARC research project funded by the European Union. The feature extraction part of the GIFT was griddified. A hospital grid concept is introduced. Grid performance was measured with a real griddified system for and several job submissions. Grid computing has the potential to help computer science researchers in medical institutions to better use an existing infrastructure. Our results show that particularly computationally–intensive tasks such as the extraction of visual features from large image databases can be per formed much faster. This allows to explore more complex feature spaces and also larger image datasets.
机译：医学图像数据的数量和种类不断增加。这些数据的数字化生产使其可用于进一步的自动分析和处理。尽管自动分析在文本领域中相当普遍，但是对医学图像进行大量，种类繁多的分析是一门相对较新的学科。计算限制通常会限制分析自动生成的大量数据的可能性。网格计算为将医院内部计算基础设施用于研究项目开辟了新的可能性。本文介绍了称为GNU图像查找工具（GIFT）的基于内容的图像检索系统的网格化。这项研究的目的是展示网格计算的潜力以及可用网格计算能力为医疗应用带来的好处。我们使用ARC（高级资源连接器）中间件来受益于通过欧盟资助的KnowARC研究项目获得的分布式计算能力。 GIFT的特征提取部分已网格化。介绍了医院网格的概念。网格性能是通过一个真正的网格化系统来测量的，并提交了多个作业。网格计算有潜力帮助医疗机构中的计算机科学研究人员更好地利用现有基础架构。我们的结果表明，特别是需要大量计算的任务，例如从大型图像数据库中提取视觉特征，可以更快地完成。这允许探索更复杂的特征空间以及更大的图像数据集。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号