Parallel analog image coding and decoding by using cellular neural networks

Mamoru Tanaka, Kenneth R. Crounse, T. Roska

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Cellular neural networks (CNNs) were used to code and decode analog images. The system consisted of differential transmitter containing internal receiver. The coder and the decoder CNNs were embedded in this area. The decoding of data utilized the dynamic current distribution. This allowed detection of original input to the coder network. Each pixel was converted to binary value through dynamic quantization. Quantization was done by halftowing. Both decoding and coding algorithms used only local image information. This design eliminated the blocking effects. Even distribution of errors contributed to the valuable elimination. Simulation experiments validated the theory underlying the CNNs.

Original languageEnglish
Pages (from-to)1387-1394
Number of pages8
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE77-A
Issue number8
Publication statusPublished - Aug 1994

Fingerprint

Cellular neural networks
Image Coding
Cellular Networks
Image coding
Decoding
Neural Networks
Analogue
Quantization
Decode
Transmitter
Simulation Experiment
Elimination
Transmitters
Receiver
Coding
Pixel
Pixels
Binary
Internal
Experiments

ASJC Scopus subject areas

  • Hardware and Architecture
  • Information Systems
  • Electrical and Electronic Engineering

Cite this

Parallel analog image coding and decoding by using cellular neural networks. / Tanaka, Mamoru; Crounse, Kenneth R.; Roska, T.

In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Vol. E77-A, No. 8, 08.1994, p. 1387-1394.

Research output: Contribution to journalArticle

@article{2010c62e121942cb89916216e6a26c05,
title = "Parallel analog image coding and decoding by using cellular neural networks",
abstract = "Cellular neural networks (CNNs) were used to code and decode analog images. The system consisted of differential transmitter containing internal receiver. The coder and the decoder CNNs were embedded in this area. The decoding of data utilized the dynamic current distribution. This allowed detection of original input to the coder network. Each pixel was converted to binary value through dynamic quantization. Quantization was done by halftowing. Both decoding and coding algorithms used only local image information. This design eliminated the blocking effects. Even distribution of errors contributed to the valuable elimination. Simulation experiments validated the theory underlying the CNNs.",
author = "Mamoru Tanaka and Crounse, {Kenneth R.} and T. Roska",
year = "1994",
month = "8",
language = "English",
volume = "E77-A",
pages = "1387--1394",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "8",

}

TY - JOUR

T1 - Parallel analog image coding and decoding by using cellular neural networks

AU - Tanaka, Mamoru

AU - Crounse, Kenneth R.

AU - Roska, T.

PY - 1994/8

Y1 - 1994/8

N2 - Cellular neural networks (CNNs) were used to code and decode analog images. The system consisted of differential transmitter containing internal receiver. The coder and the decoder CNNs were embedded in this area. The decoding of data utilized the dynamic current distribution. This allowed detection of original input to the coder network. Each pixel was converted to binary value through dynamic quantization. Quantization was done by halftowing. Both decoding and coding algorithms used only local image information. This design eliminated the blocking effects. Even distribution of errors contributed to the valuable elimination. Simulation experiments validated the theory underlying the CNNs.

AB - Cellular neural networks (CNNs) were used to code and decode analog images. The system consisted of differential transmitter containing internal receiver. The coder and the decoder CNNs were embedded in this area. The decoding of data utilized the dynamic current distribution. This allowed detection of original input to the coder network. Each pixel was converted to binary value through dynamic quantization. Quantization was done by halftowing. Both decoding and coding algorithms used only local image information. This design eliminated the blocking effects. Even distribution of errors contributed to the valuable elimination. Simulation experiments validated the theory underlying the CNNs.

UR - http://www.scopus.com/inward/record.url?scp=0028481863&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028481863&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0028481863

VL - E77-A

SP - 1387

EP - 1394

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 8

ER -