Systolic implementation of higher-order CMAC and its application in colour calibration

J. S. Ker, Yau-Hwang Kuo, Bin-Da Liu

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

The primary advantages of a CMAC neural network are fast learning and insensitivity to the order in which training patterns are presented. The authors present an extended direct weight cell address mapping mechanism based on a linear systolic array architecture to realise a higher-order CMAC neural network with digital hardware. This higher-order CMAC implementation has been applied to calibrate and compensate the nonlinearity of chromatic mapping between colour scanning and printing devices in a colour image reproduction environment. A 20MHz prototyped CMAC chip for colour calibration has been implemented to confirm the proposed design approach. Using this prototype, the authors were able to achieve reproduced colour images with rich and vivid colours which strongly resemble the original.

Original languageEnglish
Pages (from-to)129-137
Number of pages9
JournalIEE Proceedings: Circuits, Devices and Systems
Volume144
Issue number3
DOIs
Publication statusPublished - 1997 Jan 1

Fingerprint

Calibration
Color
Neural networks
Systolic arrays
Printing
Scanning
Hardware

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering

Cite this

@article{9a25a9f02253468db66d933f55ded125,
title = "Systolic implementation of higher-order CMAC and its application in colour calibration",
abstract = "The primary advantages of a CMAC neural network are fast learning and insensitivity to the order in which training patterns are presented. The authors present an extended direct weight cell address mapping mechanism based on a linear systolic array architecture to realise a higher-order CMAC neural network with digital hardware. This higher-order CMAC implementation has been applied to calibrate and compensate the nonlinearity of chromatic mapping between colour scanning and printing devices in a colour image reproduction environment. A 20MHz prototyped CMAC chip for colour calibration has been implemented to confirm the proposed design approach. Using this prototype, the authors were able to achieve reproduced colour images with rich and vivid colours which strongly resemble the original.",
author = "Ker, {J. S.} and Yau-Hwang Kuo and Bin-Da Liu",
year = "1997",
month = "1",
day = "1",
doi = "10.1049/ip-cds:19971005",
language = "English",
volume = "144",
pages = "129--137",
journal = "IET Circuits, Devices and Systems",
issn = "1751-858X",
publisher = "Institution of Engineering and Technology",
number = "3",

}

Systolic implementation of higher-order CMAC and its application in colour calibration. / Ker, J. S.; Kuo, Yau-Hwang; Liu, Bin-Da.

In: IEE Proceedings: Circuits, Devices and Systems, Vol. 144, No. 3, 01.01.1997, p. 129-137.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Systolic implementation of higher-order CMAC and its application in colour calibration

AU - Ker, J. S.

AU - Kuo, Yau-Hwang

AU - Liu, Bin-Da

PY - 1997/1/1

Y1 - 1997/1/1

N2 - The primary advantages of a CMAC neural network are fast learning and insensitivity to the order in which training patterns are presented. The authors present an extended direct weight cell address mapping mechanism based on a linear systolic array architecture to realise a higher-order CMAC neural network with digital hardware. This higher-order CMAC implementation has been applied to calibrate and compensate the nonlinearity of chromatic mapping between colour scanning and printing devices in a colour image reproduction environment. A 20MHz prototyped CMAC chip for colour calibration has been implemented to confirm the proposed design approach. Using this prototype, the authors were able to achieve reproduced colour images with rich and vivid colours which strongly resemble the original.

AB - The primary advantages of a CMAC neural network are fast learning and insensitivity to the order in which training patterns are presented. The authors present an extended direct weight cell address mapping mechanism based on a linear systolic array architecture to realise a higher-order CMAC neural network with digital hardware. This higher-order CMAC implementation has been applied to calibrate and compensate the nonlinearity of chromatic mapping between colour scanning and printing devices in a colour image reproduction environment. A 20MHz prototyped CMAC chip for colour calibration has been implemented to confirm the proposed design approach. Using this prototype, the authors were able to achieve reproduced colour images with rich and vivid colours which strongly resemble the original.

UR - http://www.scopus.com/inward/record.url?scp=0031168312&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031168312&partnerID=8YFLogxK

U2 - 10.1049/ip-cds:19971005

DO - 10.1049/ip-cds:19971005

M3 - Article

AN - SCOPUS:0031168312

VL - 144

SP - 129

EP - 137

JO - IET Circuits, Devices and Systems

JF - IET Circuits, Devices and Systems

SN - 1751-858X

IS - 3

ER -