Sequential Karhunen-Loeve basis extraction and its application to images

Author

Levy, A. ; Lindenbaum, M.

Author_Institution

Dept. of Math., Technion-Israel Inst. of Technol., Haifa, Israel

Volume

2

fYear

1998

fDate

4-7 Oct 1998

Firstpage

456

Abstract

The Karhunen-Loeve (KL) transform is an optimal method for approximating a set of vectors, which was used in image processing and computer vision for several tasks. Its computational demands and its batch calculation nature have limited its application. Here we present a new, sequential algorithm for calculating the KL basis, which is faster in typical applications and is especially advantageous for image sequences: the KL basis calculation is done with much lower delay and allows for dynamic updating of object databases for recognition. Systematic tests of the implemented algorithm show that these advantages are indeed obtained with the same accuracy available from batch KL algorithms

Keywords

Karhunen-Loeve transforms; approximation theory; computer vision; image recognition; image sequences; visual databases; accuracy; batch KL algorithms; batch calculation; computer vision; delay; dynamic updating; image processing; image recognition; image sequences; object databases; optimal method; sequential Karhunen-Loeve basis extraction; sequential algorithm; vectors approximation; Application software; Computer vision; Delay; Image databases; Image processing; Image recognition; Image sequences; Mathematics; Sparse matrices; System testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1998. ICIP 98. Proceedings. 1998 International Conference on

Conference_Location

Chicago, IL

Print_ISBN

0-8186-8821-1

Type

conf

DOI

10.1109/ICIP.1998.723422

Filename

723422