In the field of content-based retrieval of multimedia documents, 3D models have recently gained increasing attention for a number of reasons: advancements in 3D hardware and software technologies, their ever decreasing prices and increasing availability, affordable 3D authoring tools, and the establishment of such open standards for 3D data interchange as VRML and X3D.

In this paper we present a solution for content-based retrieval of 3D models that relies on a description of each model by means of a curvature map: After an initial preprocessing of the model, differential properties of points on the surface of the 3D object are evaluated; the model surface is then warped into an ellipsoid, and is mapped onto a 2D map retaining curvature information of the original model. Matching is performed by comparing the map of the query against the maps of database models. In particular, the approach expounded on in this paper exploits weighted walkthroughs to describe curvature maps. Experimental results reveal that this technique, initially developed to describe images, can be successfully applied to curvature maps.