The aim of this doctoral thesis is in the first place to establish a model that permits the comparison of Content Based Image Retrieval Systems‘ returns to given QbE queries. This model shall then be applied to enhance the result in a distributed QbE query scenario.

The first step will be to find and to categorize the current multimedia databases that allow manipulations of their ranking strategies. It is then necessary to understand what features of the multimedia data are used and how their relevance is weighted to obtain the ranking of the multimedia databases. In the following we will develop a model, using the so far obtained results, that describes how multimedia databases can be tuned, meaning if the databases had the same content, after this tuning they would answer with a similar result list. In the last step we will implement the developed model and evaluate the output and the performance.