Distributed image indexing and retrieval with mobile agents

At present, the de-facto platform for providing contents in the Internet is the World Wide Web. A technology, which is now emerging on the Web, is content-based retrieval (CBR). A content-based query matches examples or prototypes to known instances of a certain media type based on a measure of similarity. For efficiency, similarity measures are frequently computed on sets of discriminant features (so-called feature vectors) being extracted a-priori from the stored media. A number of CBR systems have already been developed. Hardly any article related to CBR fails to acknowledge IBM’s QBIC system [6] and MIT’s Photobook [8]. Both operate on images available locally, though. Smith et al. [12, 13] developed a CBR system for the World Wide Web. Scarloff et al. [11] developed a content-based image query system including a gatherer collecting images from the Web. Beigi et al. [2] already applied the principle of meta-searching to a number of available image search engines.