Merging image databases as an example for information integration

Abstract

By integrating information we mean pulling information pieces from various sources together with as little loss of information and as little redundancy as possible. Operation Research techniques applied in many situations, like heuristic algorithms, meta-heuristics, similarity measures, use of constraints, combinatorial algorithms and collaborative efforts provide valuable tools for this process, if augmented by other techniques known from knowledge management and information retrieval. Given the huge amount of information we are confronted with, information integration is one of the biggest challenges of this century. This paper describes methods and techniques for the information integration process. It is interesting to note that in areas such as information integration, consolidation and simplification that are not usually at the heart of Operation Research, some of the most important techniques used in this field can be applied successfully. Information integration is an important topic but, in general no really convincing approaches have been discovered, so far. By limiting the domain to the integration of image databases the problem becomes tractable, by using the mentioned techniques and also applying methods ranging from image processing, knowledge management and natural language processing to feature engineering.