Object identification with surface signatures

Abstract

In this paper we describe a model-based object identification system. Given a set of 3D objects and a scene containing one or more of these objects, the system identifies which objects appear in the scene by matching surface signatures. Surface signatures are statistical features which are uniform for a uniform surface. Two types of surfaces are employed; curvature signatures and spectral signatures. Furthermore, the system employs an inexpensive acquisition setup consisting of a single CCD camera and two light sources. The system has been tested on 95 observed-surfaces and 77 objects of varying degrees of curvature and color with good results.