Abstract
A medical image usually contains images of several organs, with each organ being unique. It is a prerequisite for a physician to attentively examine more than one organ, so-called organs of interest. In this paper, we present an experimental design of an intelligent Content-based Medical Image Retrieval (CBMIR)for handling multiple organs of interest through three main processes. First, CBMIR identifies all such organs by comparing the images directly using the Hausdorff distance to the single, healthy organs stored in an organ database. After that, CBMIR builds image classes using neural networks and, finally, recognizes the proper class for a query image using a multicriteria optimization approach.