Concept-based indexing and retrieval of multimedia documents

In this work, we face the problem of multimedia document indexing with reference to a specific application field, the radiological ward where automatic information management by content is an urgent need. Here, a multimedia document is composed of text and images. The keystone of the approach is the image indexing which is performed in an indirect way: the description of the image (made by an expert, in our case a physician) is further synthesised and formalised to be used by the computer. In this Paper, we propose a concept-based indexing of the description of the images which is based on Farradane's work. The basic proposal has been extended to deal with specific requirements of the considered application and to be automatically performed. A first prototype of a multimedia information retrieval system has been implemented with the goal of validating the method in the specific application.