Research On Heterogeneous Data Access In Medical Image Grid

In the current information construction process of health industry, many hospitals established a variety of medical information systems. Meanwhile, among hospitals even among different departments within the same hospital, the information systems are the mutual isolated, unable to achieve the full data sharing resources. Therefore, the establishment of an effective medical information grid platform for the development of medical services and medical research is of great significance.In order to access transparently to heterogeneous data resources, medical image grid platform heterogeneous data access system puts forward a certain of common data model and query language , with the form of gird services to provide users with the comprehensive ability of data access and integration. As the key function of information integration, heterogeneous data access system presents a SQL-like query language, and realizes the analysis, decomposition and optimization of the query language. Furthermore, establishing the global patient index and data cache mechanism improves the medical image grid data access efficiency.When the medical data resource integration is completed, users will be faced with the mass of medical information. Confronted with this situation, the establishment of an efficient and accurate information search system is an indispensable part of heterogeneous data access. In this regard, statistical language model algorithm which is based on statistical probability theory, is very suitable for massive medical data searching in medical image grid, therefore the search system use statistical language model, and makes some improvements. Using the natural language processing related technologies, puts forward the similarity calculation method of the query words and documents which includes the mutual information and statistical probability. By extracting word frequency and coexistence frequency from data set, calculates the relation between the terms with the help of knowledge dictionaries, then establishes the global association indexing.Function and performance test of system show that, heterogeneous data access system provides a transparent data access interface to upper application, and integrates the heterogeneous data in a variety of medical information system. Meanwhile, the adoption of global patients index, medical data cache and so on improves the performance of the system effectively.