Share Email Print
cover

Proceedings Paper

Building an optimal hierarchy of classification features
Author(s): Vladimir I. Klokov
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The problem of finding the most meaningful features encountered in pattern recognition applications is studied. The number of features is supposed to be large enough, while the volume of statistical data is limited. In that case the factor analysis procedures to find the most meaningful features turn to be inefficient. Therefore it is suggested to introduce the separation in the space of features equal to the Euclidean distance with positive weighting factors for each component. The zero weighting factor means that the respective feature is not used for classification. The problem posed is that of optimal selection of the weighting factors.

Paper Details

Date Published: 5 May 1999
PDF: 4 pages
Proc. SPIE 3687, International Workshop on Nondestructive Testing and Computer Simulations in Science and Engineering, (5 May 1999); doi: 10.1117/12.347428
Show Author Affiliations
Vladimir I. Klokov, St. Petersburg State Technical Univ. (Russia)


Published in SPIE Proceedings Vol. 3687:
International Workshop on Nondestructive Testing and Computer Simulations in Science and Engineering
Alexander I. Melker, Editor(s)

© SPIE. Terms of Use
Back to Top