A flexible system for the retrieval of shapes in binary images

Abstract

In this paper a flexible retrieval system of shapes present in binary digital images is described: it allows customizing the retrieval function to evaluate weighted criteria constraining distinct shape characteristics of the objects in the images such as global features of contour (represented by the Fourier Coefficients), contours irregularities (represented by the Multifractal Spectrum), presence of concavities-convexities (represented by the Contour Scale Space distribution). Further also the matching function comparing the representations of the shapes can be tuned to define a more or less strict interpretation of similarity. The evaluation experiments showed that this system can be suited to different retrieval purposes, and that generally the combination of the distinct shape indexing criteria increases both Recall and Precision with respect to the application of any single indexing criterion alone. © Springer-Verlag Berlin Heidelberg 2007.