Abstract

A framework for a large class of invariant-pattern-recognition problems, is suggested. The proposed approach combines the properties of uniqueness, registration, and optimality with the ability to use natural symmetries of input patterns. The approach is based on cross correlation. The cross-correlation function with respect to a group of transformations is defined. The properties of this function and methods of its calculation are discussed. As a special case, it is shown formally that the cross correlation with respect to the motion group is the basis of such techniques as Radon transforms and the Hough technique. The roles of symmetry and parameters are examined. A technique for decomposition of pictures into sums of separable functions associated with groups of transformations is applied to the calculation of the cross-correlation function. Several examples of applications of the proposed approach are given.

© 1989 Optical Society of America

Fourier cross correlation and invariance transformations for an optimal recognition of functions deformed by affine groups

Joseph Segman
J. Opt. Soc. Am. A 9(6) 895-902 (1992)

Scale and projection invariant pattern recognition

David Mendlovic, Naim Konforti, and Emanuel Marom
Appl. Opt. 28(23) 4982-4986 (1989)

Coordinate-transformed filter for shift-invariant and scale-invariant pattern recognition

Dan Cojoc, M. Teresa Molina, Javier García, and Carlos Ferreira
Appl. Opt. 36(20) 4812-4815 (1997)

Wavelet-transform-based composite filters for invariant pattern recognition

Zeev Zalevsky, Ido Ouzieli, and David Mendlovic
Appl. Opt. 35(17) 3141-3147 (1996)

Relationship between integral transform invariances and Lie group theory

Mario Ferraro and Terry M. Caelli
J. Opt. Soc. Am. A 5(5) 738-742 (1988)