Abstract:

In medical science information domain, researchers use the DICOM data to store medical graphic files which are created by radio examination equipment. The advantages of the DICOM standard are standardization and semantization, which provide a unified data exchange mode amongst all types of medical image equipment and medical image processing systems. A DICOM image contains rich semantic information, including patient related, examination related and image related information. However, at present, most systems make little use of its semantic information, especially in data mining. The systems store and describe image related information by constructing a relational database. This is contrary to the original intention of the organization that created the DICOM standards. The main cause of the current situation is that the Chinese system manufacturers only make use of the information exchange function of the DICOM standards, but they lack semantic understanding. In order to solve the above problems, we study information retrieval models, retrieval methods and retrieval optimization methods based on DICOM semantic information. The semantic model of DICOM standards is extended according to the preferences of current Chinese users. The text-fuzzy-query method and the data-fuzzy-query method are both used. Finally, the concept of intelligent agent for DICOM semantic retrieval is presented.

Key words: DICOM SR, semantic query, knowledge base, intelligent agent