Mean square error binary tree gain-shape vector quantization via hyperplane testing

Hiroshi Watanabe, Yoshiyuki Yashima

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

1 Citation (Scopus)

Abstract

Summary form only given, as follows. Tree-search vector quantization is an effective coding scheme at high bit rates to circumvent the high complexity of codevector search. The authors present a binary tree-search codebook design method suitable for gain/shape vector quantization. Conventional codebooks are often designed using the LBG (Linde-Buzo-Gray) algorithm when input signal distribution is unknown. In the LBG algorithm, codebook performance depends on the vectors used for the first iteration update. A splitting technique is usually used to determine the initial vectors if the input vectors have various norms or various average values. However, such a splitting technique is not available for gain/shape vector quantization, since input vectors have uniform norms and sometimes have average values of zero. When the codebook has binary tree structure, the training sequence, a set of input vectors, can be divided into two groups by using a hyperplane testing algorithm. This hyperplane is characterized by a first principal component vector. Validity of the proposed partition procedure is proved with the mean-square-error criterion.

Original languageEnglish
Title of host publicationIEEE 1988 Int Symp on Inf Theory Abstr of Pap
Place of PublicationNew York, NY, USA
PublisherPubl by IEEE
Pages163-164
Number of pages2
Volume25 n 13
Publication statusPublished - 1988
Externally publishedYes

Fingerprint

Binary trees
Vector quantization
Mean square error
Testing

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Watanabe, H., & Yashima, Y. (1988). Mean square error binary tree gain-shape vector quantization via hyperplane testing. In IEEE 1988 Int Symp on Inf Theory Abstr of Pap (Vol. 25 n 13, pp. 163-164). New York, NY, USA: Publ by IEEE.

Mean square error binary tree gain-shape vector quantization via hyperplane testing. / Watanabe, Hiroshi; Yashima, Yoshiyuki.

IEEE 1988 Int Symp on Inf Theory Abstr of Pap. Vol. 25 n 13 New York, NY, USA : Publ by IEEE, 1988. p. 163-164.

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

Watanabe, H & Yashima, Y 1988, Mean square error binary tree gain-shape vector quantization via hyperplane testing. in IEEE 1988 Int Symp on Inf Theory Abstr of Pap. vol. 25 n 13, Publ by IEEE, New York, NY, USA, pp. 163-164.
Watanabe H, Yashima Y. Mean square error binary tree gain-shape vector quantization via hyperplane testing. In IEEE 1988 Int Symp on Inf Theory Abstr of Pap. Vol. 25 n 13. New York, NY, USA: Publ by IEEE. 1988. p. 163-164
Watanabe, Hiroshi ; Yashima, Yoshiyuki. / Mean square error binary tree gain-shape vector quantization via hyperplane testing. IEEE 1988 Int Symp on Inf Theory Abstr of Pap. Vol. 25 n 13 New York, NY, USA : Publ by IEEE, 1988. pp. 163-164
@inproceedings{4365be2f66bb47e39f750bc13c59f32f,
title = "Mean square error binary tree gain-shape vector quantization via hyperplane testing",
abstract = "Summary form only given, as follows. Tree-search vector quantization is an effective coding scheme at high bit rates to circumvent the high complexity of codevector search. The authors present a binary tree-search codebook design method suitable for gain/shape vector quantization. Conventional codebooks are often designed using the LBG (Linde-Buzo-Gray) algorithm when input signal distribution is unknown. In the LBG algorithm, codebook performance depends on the vectors used for the first iteration update. A splitting technique is usually used to determine the initial vectors if the input vectors have various norms or various average values. However, such a splitting technique is not available for gain/shape vector quantization, since input vectors have uniform norms and sometimes have average values of zero. When the codebook has binary tree structure, the training sequence, a set of input vectors, can be divided into two groups by using a hyperplane testing algorithm. This hyperplane is characterized by a first principal component vector. Validity of the proposed partition procedure is proved with the mean-square-error criterion.",
author = "Hiroshi Watanabe and Yoshiyuki Yashima",
year = "1988",
language = "English",
volume = "25 n 13",
pages = "163--164",
booktitle = "IEEE 1988 Int Symp on Inf Theory Abstr of Pap",
publisher = "Publ by IEEE",

}

TY - GEN

T1 - Mean square error binary tree gain-shape vector quantization via hyperplane testing

AU - Watanabe, Hiroshi

AU - Yashima, Yoshiyuki

PY - 1988

Y1 - 1988

N2 - Summary form only given, as follows. Tree-search vector quantization is an effective coding scheme at high bit rates to circumvent the high complexity of codevector search. The authors present a binary tree-search codebook design method suitable for gain/shape vector quantization. Conventional codebooks are often designed using the LBG (Linde-Buzo-Gray) algorithm when input signal distribution is unknown. In the LBG algorithm, codebook performance depends on the vectors used for the first iteration update. A splitting technique is usually used to determine the initial vectors if the input vectors have various norms or various average values. However, such a splitting technique is not available for gain/shape vector quantization, since input vectors have uniform norms and sometimes have average values of zero. When the codebook has binary tree structure, the training sequence, a set of input vectors, can be divided into two groups by using a hyperplane testing algorithm. This hyperplane is characterized by a first principal component vector. Validity of the proposed partition procedure is proved with the mean-square-error criterion.

AB - Summary form only given, as follows. Tree-search vector quantization is an effective coding scheme at high bit rates to circumvent the high complexity of codevector search. The authors present a binary tree-search codebook design method suitable for gain/shape vector quantization. Conventional codebooks are often designed using the LBG (Linde-Buzo-Gray) algorithm when input signal distribution is unknown. In the LBG algorithm, codebook performance depends on the vectors used for the first iteration update. A splitting technique is usually used to determine the initial vectors if the input vectors have various norms or various average values. However, such a splitting technique is not available for gain/shape vector quantization, since input vectors have uniform norms and sometimes have average values of zero. When the codebook has binary tree structure, the training sequence, a set of input vectors, can be divided into two groups by using a hyperplane testing algorithm. This hyperplane is characterized by a first principal component vector. Validity of the proposed partition procedure is proved with the mean-square-error criterion.

UR - http://www.scopus.com/inward/record.url?scp=0024124653&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0024124653&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0024124653

VL - 25 n 13

SP - 163

EP - 164

BT - IEEE 1988 Int Symp on Inf Theory Abstr of Pap

PB - Publ by IEEE

CY - New York, NY, USA

ER -