Fast and efficient multi-layer CNN-UM emulator using FPGA

Zoltán Nagy, P. Szolgay

Research output: Contribution to journalArticle

Abstract

A new emulated digital multi-layer CNN-UM chip architecture called Falcon has been developed. Simulation running time can be hundred times shorter using the Falcon processor array compared to the software simulation. This huge computing power makes real time image processing possible. In this paper the main steps of the FPGA implementation and optimization are introduced. The Distributed Arithmetic technique is used to optimize the architecture on FPGAs. Using this technique, smaller and faster arithmetic units can be designed than using conventional approach where multiplier cores and adder trees are used to compute the state equation of the CNN array.

Original languageEnglish
Pages (from-to)57-70
Number of pages14
JournalPeriodica Polytechnica, Electrical Engineering
Volume47
Issue number1-2
Publication statusPublished - 2003

Fingerprint

Field programmable gate arrays (FPGA)
Adders
Parallel processing systems
Equations of state
Image processing

Keywords

  • Cellular neural networks
  • Reconfigurable computing

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Fast and efficient multi-layer CNN-UM emulator using FPGA. / Nagy, Zoltán; Szolgay, P.

In: Periodica Polytechnica, Electrical Engineering, Vol. 47, No. 1-2, 2003, p. 57-70.

Research output: Contribution to journalArticle

@article{e23c9f6de0174be29f9251f30195bc2d,
title = "Fast and efficient multi-layer CNN-UM emulator using FPGA",
abstract = "A new emulated digital multi-layer CNN-UM chip architecture called Falcon has been developed. Simulation running time can be hundred times shorter using the Falcon processor array compared to the software simulation. This huge computing power makes real time image processing possible. In this paper the main steps of the FPGA implementation and optimization are introduced. The Distributed Arithmetic technique is used to optimize the architecture on FPGAs. Using this technique, smaller and faster arithmetic units can be designed than using conventional approach where multiplier cores and adder trees are used to compute the state equation of the CNN array.",
keywords = "Cellular neural networks, Reconfigurable computing",
author = "Zolt{\'a}n Nagy and P. Szolgay",
year = "2003",
language = "English",
volume = "47",
pages = "57--70",
journal = "Periodica polytechnica Electrical engineering and computer science",
issn = "2064-5260",
publisher = "Budapest University of Technology and Economics",
number = "1-2",

}

TY - JOUR

T1 - Fast and efficient multi-layer CNN-UM emulator using FPGA

AU - Nagy, Zoltán

AU - Szolgay, P.

PY - 2003

Y1 - 2003

N2 - A new emulated digital multi-layer CNN-UM chip architecture called Falcon has been developed. Simulation running time can be hundred times shorter using the Falcon processor array compared to the software simulation. This huge computing power makes real time image processing possible. In this paper the main steps of the FPGA implementation and optimization are introduced. The Distributed Arithmetic technique is used to optimize the architecture on FPGAs. Using this technique, smaller and faster arithmetic units can be designed than using conventional approach where multiplier cores and adder trees are used to compute the state equation of the CNN array.

AB - A new emulated digital multi-layer CNN-UM chip architecture called Falcon has been developed. Simulation running time can be hundred times shorter using the Falcon processor array compared to the software simulation. This huge computing power makes real time image processing possible. In this paper the main steps of the FPGA implementation and optimization are introduced. The Distributed Arithmetic technique is used to optimize the architecture on FPGAs. Using this technique, smaller and faster arithmetic units can be designed than using conventional approach where multiplier cores and adder trees are used to compute the state equation of the CNN array.

KW - Cellular neural networks

KW - Reconfigurable computing

UR - http://www.scopus.com/inward/record.url?scp=2942601181&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=2942601181&partnerID=8YFLogxK

M3 - Article

VL - 47

SP - 57

EP - 70

JO - Periodica polytechnica Electrical engineering and computer science

JF - Periodica polytechnica Electrical engineering and computer science

SN - 2064-5260

IS - 1-2

ER -