Medical Image Storage and Retrieval Strategies in the LUX-IMACS project

David Bandon, Robert Kanz, Valerie Boissart, Chantal Debas, Gerard Evers, Jacques Duchêne, Claude Wehenkel
1994 Open research Areas in Information Retrieval  
The medical image management for Picture Archiving and Communication Systems (PACS) requires a high-speed and large storage system because of image volumes up to 50 megabytes. For instance, the PACS system developed at the Centre Hospitalier de Luxembourg in the LUX-IMACS project will have to handle 7 terabytes of image data. To solve the various performance problems related to this huge volume, an image server implementing a distributed database, a centralized image bank and hierarchical
more » ... e levels was designed in the LUX-IMACS project. The database is a common information system for the radiological department by integrating RIS (Radiology Information System) and PACS systems in one data model. This approach increases coherence of data model and enhances the development of storage methods. Images migrate inside hierarchical storage according to their retrieval probability. This probability, based on the medical relevance of an image and the organization model, is expected to improve the work flow control. First studies characterized the image relevance and analyzed the current medical scenarios of working. The PACS development in the Centre Hospitalier de Luxembourg is performed by a modular approach. Thus, a PACS kernel including image acquisition, consultation and archiving was already implemented on the basis of the designed architecture. In this system, images migrate in a two levels hierarchical storage, i.e. short-term and long-term memories, for archiving, retrieval or memory purge. We plan to optimize in the future this storage via organization model and image relevance.
dblp:conf/riao/BandonKBDEDW94 fatcat:fzh3vh4cnrfb3gygpvym4szr6u