Masud Mehedi, Hossain M Shamim, Alamri Atif
IEEE Trans Inf Technol Biomed. 2012 Nov;16(6):1015-23. doi: 10.1109/TITB.2012.2202244. Epub 2012 Jun 1.
E-Health systems provide a collaborative platform for sharing patients medical data typically stored in distributed autonomous healthcare data sources. Each autonomous source stores its medical and multimedia data without following any global structure. This causes heterogeneity in the underlying sources with respect to the data and storage structure. Therefore, a data interoperability mechanism is required for sharing the data among the heterogeneous sources. A proper metadata structure is also necessary to represent multimedia content in the sources to enable efficient query processing. Considering these needs, we present an interoperability solution for sharing data among heterogeneous data sources. We also propose a metadata management framework for medical multimedia con-tent including X-ray, ECG, MRI, and ultrasound images. The framework identifies features, generates and represents metadata, and produces identifiers for the medical multimedia content to facilitate efficient query processing. The framework has been tested with various user queries and the accuracy of the query results evaluated by means of precision, recall, and user feedback methods. The results confirm the effectiveness of the proposed approach.
电子健康系统提供了一个协作平台，用于共享通常存储在分布式自主医疗数据源中的患者医疗数据。每个自主数据源存储其医疗和多媒体数据时不遵循任何全局结构。这导致基础数据源在数据和存储结构方面存在异构性。因此，需要一种数据互操作性机制来在异构数据源之间共享数据。还需要一个适当的元数据结构来表示源中的多媒体内容，以实现高效的查询处理。考虑到这些需求，我们提出了一种在异构数据源之间共享数据的互操作性解决方案。我们还提出了一个用于医疗多媒体内容（包括X射线、心电图、磁共振成像和超声图像）的元数据管理框架。该框架识别特征、生成并表示元数据，并为医疗多媒体内容生成标识符，以促进高效的查询处理。该框架已通过各种用户查询进行测试，并通过精确率、召回率和用户反馈方法评估查询结果的准确性。结果证实了所提方法的有效性。
