Intraframe coding with DCT-VQ for image sequence compression

Author

Türker, Mustafii Ali ; Severcan, Mete

Author_Institution

Ankara Electron. Res. & Dev. Inst.., Turkey

fYear

1994

fDate

12-14 Apr 1994

Firstpage

238

Abstract

A VQ codebook generation technique is introduced for intraframe image sequence compression. It involves some stochastic methods applied to the classical Kohonen learning algorithm, so that optimal codebooks can be generated without being trapped in poor local minima of MSE distortion function. Codebook generation time is much faster than other stochastic methods. Using the technique in a vector quantizer, following a discrete cosine transform, high PSNR performance is obtained at 0.25 bpp encoded frames

Keywords

discrete cosine transforms; image coding; image sequences; self-organising feature maps; stochastic processes; transform coding; vector quantisation; DCT-VQ; MSE distortion function; classical Kohonen learning algorithm; codebook generation technique; codebook generation time; discrete cosine transform; high PSNR performance; image sequence compression; intraframe coding; optimal codebooks; poor local minima; stochastic methods; vector quantizer; Councils; Data compression; Discrete cosine transforms; Electronic mail; Image coding; Image sequences; Karhunen-Loeve transforms; Research and development; Stochastic processes; Transform coding;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrotechnical Conference, 1994. Proceedings., 7th Mediterranean

Conference_Location

Antalya

Print_ISBN

0-7803-1772-6

Type

conf

DOI

10.1109/MELCON.1994.380924

Filename

380924