A NEURAL-NETWORK POSITION-INDEPENDENT MULTIPLE PATTERN RECOGNISER

Authors
GRIMES C PICTON PD ELLIMAN DG
Citation
C. Grimes et al., A NEURAL-NETWORK POSITION-INDEPENDENT MULTIPLE PATTERN RECOGNISER, Artificial intelligence in engineering, 10(2), 1996, pp. 117-126
Citations number
7
Categorie Soggetti
Computer Application, Chemistry & Engineering","Computer Science Artificial Intelligence",Engineering
Journal title
Artificial intelligence in engineering → ACNP
ISSN journal
09541810
Volume
10
Issue
2
Year of publication
1996
Pages
117 - 126
Database
ISI
SICI code
0954-1810(1996)10:2<117:ANPMPR>2.0.ZU;2-U
Abstract
This paper describes a neural network model for computer vision which has position invariant properties. The network is designed to form par t of a more comprehensive vision system. The purpose of the network is to classify features in a position independent manner and retain the spatial relationship between detected features. Inherent parallelism i n the network allows multiple features to be simultaneously classified with the spatial relationships preserved.