| A Mobile Multi-Agent Image Retrieval System with applications to mammography is presented. The image retrieval strategy suggested uses the inherent strengths of Mobile Agents to return the best possible matches from a distributed set of images in Digital Medical Image Libraries available on the Internet. The proposed approach is tested using a simulation of the Digital Medical Image Library. We study the dynamics of the system from the perspective of parallel processing, intelligent pre-processing, and intelligent-communicating Agents. In this work we focus on various strategies which agents can implement, and examine the efficiency of these strategies. While the accuracy of the system is important, we assume that the accuracy comes from the quality of the returned results. Thus our emphasis is on the rate at which good results can be returned. Our contribution is the creation of these competing strategies, which are unique to this problem domain, and which are quantitatively evaluated in term of the rate at which good results are returned. |
