Scientific article
Open access

Invariance Signatures: Characterizing contours by their departures from invariance

Published inComputer vision and image understanding, vol. 77, no. 3, p. 284-316
Publication date2000

In this paper, a new invariant feature of two-dimensional contours is reported: the Invariance Signature. The Invariance Signature is a measure of the degree to which a contour is invariant under a variety of transformations, derived from the theory of Lie transformation groups. It is shown that the Invariance Signature is itself invariant under shift, rotation and scaling of the contour. Since it is derived from local properties of the contour, it is well-suited to a neural network implementation. It is shown that a Model-Based Neural Network (MBNN) can be constructed which computes the Invariance Signature of a contour, and classifies patterns on this basis. Experiments demonstrate that Invariance Signature networks can be employed successfully for shift-, rotation- and scale-invariant optical character recognition.

Citation (ISO format)
SQUIRE, David, CAELLI, Terry M. Invariance Signatures: Characterizing contours by their departures from invariance. In: Computer vision and image understanding, 2000, vol. 77, n° 3, p. 284–316. doi: 10.1006/cviu.2000.0809
Main files (1)
Article (Published version)
ISSN of the journal1077-3142

Technical informations

Creation03/03/2015 4:36:16 PM
First validation03/03/2015 4:36:16 PM
Update time03/14/2023 10:57:13 PM
Status update03/14/2023 10:57:12 PM
Last indexation01/16/2024 5:04:01 PM
All rights reserved by Archive ouverte UNIGE and the University of GenevaunigeBlack