Towards Adaptive Ontology-Based Image Retrieval

Since the use of large image databases gains in importance nowadays, efficient querying and browsing through image repositories becomes increasingly essential. Compared to text retrieval techniques there are even more problems with image retrieval. Particularly, the semantic gap between low-level visual features of images and high-level human perception of inferred semantic contents decreases the performance of traditional content-based image retrieval systems. The first important step for the correlation of image data with cognitive processes is the identification of discriminative features in the data. The next decisive step is to extract high-level knowledge from this data in order to provide a confident interpretation of signals into symbols. In this paper we demonstrate our first conceptual notions about the integration of spatial context and semantic concepts into the feature extraction and retrieval process using the relevance feedback procedure.