Sketch based and context sensitive coding of Laplacian pyramids

Author

Vincent, John M.

Author_Institution

Dept. of Electron. Syst. Eng., Essex Univ., Colchester, UK

fYear

1988

fDate

11-14 Apr 1988

Firstpage

804

Abstract

An improvement in pyramidal coding is described which goes beyond first-order statistics by exploiting structural redundancies within Laplacian pyramids. Their three-dimensional structures are described with a novel sketch pyramid, which is used in a new context-sensitive method for coding Laplacian pyramids. This approach is enhanced by using the Gaussian function as the generating kernel, and the sinc function for the expansion filter. The compression ratios obtained from pyramidal coding are improved by a factor ranging from around 2.5 to 4, depending on scene complexity. Example results are given

Keywords

data compression; encoding; filtering and prediction theory; picture processing; Gaussian function; Laplacian pyramids; compression ratios; context sensitive coding; expansion filter; generating kernel; image coding; pyramidal coding; sinc function; sketch based coding; sketch pyramid; structural redundancies; three-dimensional structures; Decoding; Filters; Image coding; Image reconstruction; Kernel; Laplace equations; Pixel; Shape; Statistics; Systems engineering and theory;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1988. ICASSP-88., 1988 International Conference on

Conference_Location

New York, NY

ISSN

1520-6149

Type

conf

DOI

10.1109/ICASSP.1988.196707

Filename

196707