Summary: On the Existence of Complete Invariant Feature Spaces in Pattern
Recognition \Lambda
Hanns Schulz­Mirbach
Technische Universit¨at Hamburg­Harburg
Institut f¨ur Technische Informatik I
Postfach 90 10 52, 2100 Hamburg 90, Germany
Abstract
This paper presents a general and model indepen­
dent analysis of the problem of feature extraction in
pattern recognition. We will derive two criteria which
ensure the existence of a complete feature space. This
is a space which contains exactly the information rel­
evant for the classification process following feature
extraction. We discuss several possibilities for the con­
struction of such a complete feature space and present
experimental results which indicate the potential of
the proposed methods for practical applications.
1 Introduction
Pattern classification is scarcely based directly on
the information gained by some sort of sensor (i.g. a

Source: Albert-Ludwigs-Universität Freiburg, Institut für Informatik,, Lehrstuhls für Mustererkennung und Bildverarbeitung

Collections: Computer Technologies and Information Sciences