Can the sample being transmitted be used to refine its own PDF estimate?

Author

Florêncio, Dinei A. ; Simard, Patrice

Author_Institution

Microsoft Res., Redmond, WA, USA

fYear

2003

fDate

25-27 March 2003

Firstpage

233

Lastpage

242

Abstract

Many image coders map the input image to a transform domain, and encode coefficients in that domain. Usually, encoders use previously transmitted samples to help estimate the probabilities for the next sample to be encoded. This backward probability estimation provides a better PDF estimation, without need to send any side information. A new method of encoding was proposed that goes one step further: besides past samples, it also uses information about the current sample in computing the PDF for the current sample. Yet, no side information was transmitted. The initial PDF estimate is based on a Tarp filter, but probabilities are then progressively refined for non-zero samples. Results are superior to JPEG2000, and to bit-plane Tarp.

Keywords

IIR filters; image coding; image sampling; prediction theory; probability; transform coding; wavelet transforms; JPEG2000; PDF estimation; Tarp filter; backward probability estimation; bit plane Tarp; encoders; image coders; nonzero samples; probability distribution function; sample probabilites; side information; transform domain; transmitted samples; Computational complexity; Data compression; Decoding; Encoding; Entropy coding; Filters; Frequency; Probability distribution; Wavelet coefficients; Wavelet transforms;

fLanguage

English

Publisher

ieee

Conference_Titel

Data Compression Conference, 2003. Proceedings. DCC 2003

ISSN

1068-0314

Print_ISBN

0-7695-1896-6

Type

conf

DOI

10.1109/DCC.2003.1194014

Filename

1194014