Binary PACT

Author

Yoshii, Hiroto

Author_Institution

Canon Inc., Kawasaki, Japan

Volume

4

fYear

1996

fDate

25-29 Aug 1996

Firstpage

606

Abstract

The pyramid architecture classification tree (PACT) is a novel pattern recognition algorithm which has good capabilities. PACT is a kind of decision tree classifier, though the algorithm is motivated from quite different backgrounds from conventional pattern recognition algorithms. Moreover, PACT motivates us to propose a new hypothesis “a decision region in the feature space having fractal characteristics”. A theoretical model of PACT, a random cantor set problem, is proposed, and using the problem we argue that the hypothesis requires “iterative feature extraction”, which is the key point of PACT and what the conventional pattern recognition algorithms lack. A binary PACT is proposed for an implicit evidence of the hypothesis

Keywords

computer vision; decision theory; feature extraction; fractals; image matching; iterative methods; multilayer perceptrons; trees (mathematics); binary PACT; decision tree classifier; feature space; fractal characteristics; iterative feature extraction; multilayer perceptrons; pattern recognition algorithm; pyramid architecture classification tree; random cantor set; Classification tree analysis; Decision trees; Feature extraction; Fractals; Humans; Iterative algorithms; Multilayer perceptrons; Neural networks; Partitioning algorithms; Pattern recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Pattern Recognition, 1996., Proceedings of the 13th International Conference on

Conference_Location

Vienna

ISSN

1051-4651

Print_ISBN

0-8186-7282-X

Type

conf

DOI

10.1109/ICPR.1996.547636

Filename

547636