Abstract
One goal of artificial intelligence research into case-based reasoning (CBR) systems is to develop approaches for designing useful and practical interactive case-based environments. Explaining each step of the design of the case-base and of the retrieval process is critical for the application of case-based systems to the real world. We describe herein our approach to the design of IDEM--Images and Diagnosis from Examples in Medicine--a medical image case-based retrieval system for pathologists. Our approach is based on the expressiveness of an object-oriented modeling language standard: the Unified Modeling Language (UML). We created a set of diagrams in UML notation illustrating the steps of the CBR methodology we used. The key aspect of this approach was selecting the relevant objects of the system according to user requirements and making visualization of cases and of the components of the case retrieval process. Further evaluation of the expressiveness of the design document is required but UML seems to be a promising formalism, improving the communication between the developers and users.
Full text
PDF
Images in this article
Selected References
These references are in PubMed. This may not be the complete list of references from this article.
- Kahn C. E., Jr, Anderson G. M. Case-based reasoning and imaging procedure selection. Invest Radiol. 1994 Jun;29(6):643–647. doi: 10.1097/00004424-199406000-00009. [DOI] [PubMed] [Google Scholar]