Histogram-based color signatures for image indexing

Content-based image retrieval primarily used color distributions as descriptors of the image content. In order to overcome the many limitations of the description by a firstorder distribution, several higher- order distributions have been introduced. Although they can perform better, their computational complexity is prohibitive. We propose to upgrade color histogram by embedding for each color additional information about its perceptual or local statistical relevance. A second fam- ily of descriptor are introduce: histograms are parametric descriptors that allow scale-zoom tuning of histogram values. We prove that the new color distribution families are robust and easy to compute and provide a superior retrieval performance.