A fixed-length compression method using the Karhunen-Loeve transform

Author

Nasiopoulos, Panos ; Yedlin, Matthew ; Ward, Rabab K.

Author_Institution

Dept. of Electr. Eng., British Columbia Univ., Vancouver, BC, Canada

fYear

1995

fDate

17-19 May 1995

Firstpage

581

Lastpage

584

Abstract

A compression method is presented which combines the distinct advantage of being fixed length with the visual picture quality of transform-based coding methods. This method compresses the Karhunen-Loeve transform coefficients of 8×8 pixel blocks of the picture and attains a compression ratio of 15.37:1 for color images. Compared to the absolute moment block truncation coding, the other known fixed length compression method, our method improves the compression ratio by 68% and significantly improves the visual quality of the reconstructed picture

Keywords

data compression; image coding; image reconstruction; transform coding; transforms; 8×8 pixel blocks; Karhunen-Loeve transform; absolute moment block truncation coding; color images; compression ratio; fixed-length compression method; reconstructed picture; transform-based coding methods; visual picture quality; Color; Communications technology; Covariance matrix; Discrete cosine transforms; Image coding; Image reconstruction; Karhunen-Loeve transforms; Pixel; Symmetric matrices; Transform coding;

fLanguage

English

Publisher

ieee

Conference_Titel

Communications, Computers, and Signal Processing, 1995. Proceedings., IEEE Pacific Rim Conference on

Conference_Location

Victoria, BC

Print_ISBN

0-7803-2553-2

Type

conf

DOI

10.1109/PACRIM.1995.519598

Filename

519598