An improved vector quantizer design method: The codebook reorganization algorithm

Ting Wei Hou, Yuan Tsung Chen, Houng O. Ku

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Generalized Lloyd Algorithm(GLA) is important in vector quantizer design. It runs fast, but it is sensitive to initial conditions and it may find a local optimum. We propose an improved approach based on GLA, named vector quantized codebook reorganization algorithm (VQCRA). VQCRA finds better codebooks in less time, and is insensitive to initial conditions as compared with GLA.. The experiments are performed on the 10 images extracted from USC-SIPI Image Database. The performance improvement ranges from 7.3% (the size of the codebook is 32) to 46.4% (the size of the codebook is 512).

Original languageEnglish
Title of host publicationProceedings of the 9th Joint Conference on Information Sciences, JCIS 2006
DOIs
Publication statusPublished - 2006 Dec 1
Event9th Joint Conference on Information Sciences, JCIS 2006 - Taiwan, ROC, Taiwan
Duration: 2006 Oct 82006 Oct 11

Publication series

NameProceedings of the 9th Joint Conference on Information Sciences, JCIS 2006
Volume2006

Other

Other9th Joint Conference on Information Sciences, JCIS 2006
CountryTaiwan
CityTaiwan, ROC
Period06-10-0806-10-11

Fingerprint

Experiments

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Hou, T. W., Chen, Y. T., & Ku, H. O. (2006). An improved vector quantizer design method: The codebook reorganization algorithm. In Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006 [CVPRIP-52] (Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006; Vol. 2006). https://doi.org/10.2991/jcis.2006.264
Hou, Ting Wei ; Chen, Yuan Tsung ; Ku, Houng O. / An improved vector quantizer design method : The codebook reorganization algorithm. Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006. 2006. (Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006).
@inproceedings{dbe15399a8af4bb18a2974eb22683d28,
title = "An improved vector quantizer design method: The codebook reorganization algorithm",
abstract = "Generalized Lloyd Algorithm(GLA) is important in vector quantizer design. It runs fast, but it is sensitive to initial conditions and it may find a local optimum. We propose an improved approach based on GLA, named vector quantized codebook reorganization algorithm (VQCRA). VQCRA finds better codebooks in less time, and is insensitive to initial conditions as compared with GLA.. The experiments are performed on the 10 images extracted from USC-SIPI Image Database. The performance improvement ranges from 7.3{\%} (the size of the codebook is 32) to 46.4{\%} (the size of the codebook is 512).",
author = "Hou, {Ting Wei} and Chen, {Yuan Tsung} and Ku, {Houng O.}",
year = "2006",
month = "12",
day = "1",
doi = "10.2991/jcis.2006.264",
language = "English",
isbn = "9078677015",
series = "Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006",
booktitle = "Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006",

}

Hou, TW, Chen, YT & Ku, HO 2006, An improved vector quantizer design method: The codebook reorganization algorithm. in Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006., CVPRIP-52, Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006, vol. 2006, 9th Joint Conference on Information Sciences, JCIS 2006, Taiwan, ROC, Taiwan, 06-10-08. https://doi.org/10.2991/jcis.2006.264

An improved vector quantizer design method : The codebook reorganization algorithm. / Hou, Ting Wei; Chen, Yuan Tsung; Ku, Houng O.

Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006. 2006. CVPRIP-52 (Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006; Vol. 2006).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - An improved vector quantizer design method

T2 - The codebook reorganization algorithm

AU - Hou, Ting Wei

AU - Chen, Yuan Tsung

AU - Ku, Houng O.

PY - 2006/12/1

Y1 - 2006/12/1

N2 - Generalized Lloyd Algorithm(GLA) is important in vector quantizer design. It runs fast, but it is sensitive to initial conditions and it may find a local optimum. We propose an improved approach based on GLA, named vector quantized codebook reorganization algorithm (VQCRA). VQCRA finds better codebooks in less time, and is insensitive to initial conditions as compared with GLA.. The experiments are performed on the 10 images extracted from USC-SIPI Image Database. The performance improvement ranges from 7.3% (the size of the codebook is 32) to 46.4% (the size of the codebook is 512).

AB - Generalized Lloyd Algorithm(GLA) is important in vector quantizer design. It runs fast, but it is sensitive to initial conditions and it may find a local optimum. We propose an improved approach based on GLA, named vector quantized codebook reorganization algorithm (VQCRA). VQCRA finds better codebooks in less time, and is insensitive to initial conditions as compared with GLA.. The experiments are performed on the 10 images extracted from USC-SIPI Image Database. The performance improvement ranges from 7.3% (the size of the codebook is 32) to 46.4% (the size of the codebook is 512).

UR - http://www.scopus.com/inward/record.url?scp=33847720197&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33847720197&partnerID=8YFLogxK

U2 - 10.2991/jcis.2006.264

DO - 10.2991/jcis.2006.264

M3 - Conference contribution

AN - SCOPUS:33847720197

SN - 9078677015

SN - 9789078677017

T3 - Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006

BT - Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006

ER -

Hou TW, Chen YT, Ku HO. An improved vector quantizer design method: The codebook reorganization algorithm. In Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006. 2006. CVPRIP-52. (Proceedings of the 9th Joint Conference on Information Sciences, JCIS 2006). https://doi.org/10.2991/jcis.2006.264