Abstract: SIMBA is a content based image retrieval system performing queries based on image appearance. SIMBA uses two feature extraction techniques: color and texture feature extraction, provided by Content Based Image Retrieval System. In SIMBA, object position is irrelevant for image similarity hence invariant features are used for the purpose. Based on general construction method, i.e. integration over the transformation group, the invariant feature histograms are derived that catch different cues of image content: features that are strongly influenced by colour and textural features that are robust to illumination changes. The feature extraction does not require any manual interaction, so that it might be used for fully automatic annotation in heavily fluctuating image databases.

Keywords: SIMBA, invariant features, content based image retrieval, colour and texture features.