Free-Form 3-D Object Retrieval from Arbitrary Viewpoints

Abstract

This chapter addresses the problem of retrieval of free-form 3-D objects using arbitrary views of those objects as queries. It also addresses the issue of automatic selection of the best and the optimum number of views for each object in multi-view 3-D object representation and retrieval. The object boundary in each view is a 2-D shape and is represented by the maxima of its CSS image contours. After representing each object by an optimal number of views, an unknown object is recognised using a single image taken from an arbitrary viewpoint using the same shape representation.