The effect of distortion on the MDL model

Author

Gronich, Yoram ; Zamir, Ram

Author_Institution

Dept. of Electr. Eng.-Syst., Tel Aviv Univ., Israel

fYear

2001

fDate

6/23/1905 12:00:00 AM

Firstpage

156

Lastpage

157

Abstract

We investigate the consequences of lossy compression, i.e., description with distortion, on the model selection of the minimum description length (MDL) criterion. Our basic observation is that for a finite data sequence and sufficiently large distortion, a two-stage universal lossy encoder tends to under-estimate the model order of the source. We demonstrate this property by examining the behavior of a two-stage universal lossy encoder, based on pre/post-filtered entropy-coded dithered quantization, over some parametric classes of stationary Gaussian sources

Keywords

data compression; entropy codes; quantisation (signal); rate distortion theory; source coding; MDL model; distortion effect; entropy-coded dithered quantization; finite data sequence; lossy compression; minimum description length criterion; pre/post-filtered quantization; rate-distortion performance; stationary Gaussian sources; two-stage universal lossy encoder; Decoding; Distortion measurement; Electronic mail; Encoding; Integrated circuit modeling; Lagrangian functions; Parametric statistics; Probability distribution; Quantization; Rate-distortion;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Theory Workshop, 2001. Proceedings. 2001 IEEE

Conference_Location

Cairns, Qld.

Print_ISBN

0-7803-7119-4

Type

conf

DOI

10.1109/ITW.2001.955169

Filename

955169