| | |
Summary: A New Signature-Based Method for Efficient 3-D Object Recognition
Salvador Ruiz-CorreaÝ
, Linda G. ShapiroÝ
and Marina MeliaÞ
ÝDepartment of Electrical Engineering
ÞDepartment of Statistics
University of Washington, Seattle, WA 98105
sruiz@u, shapiro@cs, mmp@stat .washington.edu
Abstract
This paper considers the problem of shape-based recog-
nition and pose estimation of 3-D free-form objects in
scenes that contain occlusion and clutter. Our approach
is based on a novel set of discriminating descriptors called
spherical spin images, which encode the shape information
conveyed by classes of distributions of surface points con-
structed with respect to reference points on the surface of an
object. The key to this approach is the relationship that ex-
ists between the о metric, which compares Ò-dimensional
signatures in Euclidean space, and the metric of the com-
pact space on which the class representatives (spherical
|
