Vector quantization using matrix decompositions of codebooks for image coding

Author

Yang, Jar-Ferr ; Lu, Chiou-Liang ; Chin, Po-Wen

Author_Institution

Dept. of Electr. Eng., Nat. Cheng Kung Univ., Tainan, Taiwan

fYear

1991

fDate

11-14 Jun 1991

Firstpage

308

Abstract

Methods of using vector quantization combined with singular value decomposition (SVD) and simplified least square estimation (LSE) compensation are proposed. The complexity of computation is decreased by determining a simpler scheme to obtain the same optimal estimated singular values as the LSE. With the assistance of the matrix-decomposed codebooks, edge degradation and compression rate can be improved by using a few least-square-compensated singular values. Avoiding the high computational complexity of SVD during the encoding and decoding procedures and achieving good quality of image coding are the main contributions of the present work

Keywords

computational complexity; computerised picture processing; data compression; encoding; compression rate; computational complexity reduction; decoding procedures; edge degradation; encoding; image coding; least-square-compensated singular values; matrix decompositions; matrix-decomposed codebooks; simplified least square estimation; singular value decomposition; vector quantization; Degradation; Delta modulation; HDTV; ISDN; Image coding; Image storage; Matrix decomposition; Pulse modulation; Singular value decomposition; Vector quantization;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 1991., IEEE International Sympoisum on

Print_ISBN

0-7803-0050-5

Type

conf

DOI

10.1109/ISCAS.1991.176335

Filename

176335