Using fractal encoding for image indexing

As already documented in the literature, fractal image encoding can be exploited to obtain good image indexing systems that are able to deal with the images in compressed form, which makes them suitable for use with large databases. The system proposed in this paper has the additional interesting property of being invariant under two classes of pixel intensity transformations. This property makes the system robust with respect to several types of image transformations that can happen in practical applications: the images can be retrieved even in the presence of illumination and/or color alterations. The system can be queried via both a query by sketch and a query by example paradigms.