A SIMD neural network processor for image processing

Dongsun Kim, Hyunsik Kim, Hongsik Kim, Gunhee Han, Duckjin Chung

Research output: Contribution to journalConference article

7 Citations (Scopus)

Abstract

Artificial Neural Networks (ANNs) and image processing requires massively parallel computation of simple operator accompanied by heavy memory access. Thus, this type of operators naturally maps onto Single Instruction Multiple Data (SIMD) stream parallel processing with distributed memory. This paper proposes a high performance neural network processor whose function can be changed by programming. The proposed processor is based on the SIMD architecture that is optimized for neural network and image processing. The proposed processor supports 24 instructions, and consists of 16 Processing Units (PUs) per chip. Each PU includes 24-bit 2K-word Local Memory (LM) and a Processing Element (PE). The proposed architecture allows multichip expansion that minimizes chip-to-chip communication bottleneck. The proposed processor is verified with FPGA implementation and the functionality is verified with character recognition application.

Original languageEnglish
Pages (from-to)665-672
Number of pages8
JournalLecture Notes in Computer Science
Volume3497
Issue numberII
Publication statusPublished - 2005 Sep 26
EventSecond International Symposium on Neural Networks: Advances in Neural Networks - ISNN 2005 - Chongqing, China
Duration: 2005 May 302005 Jun 1

Fingerprint

Network Processor
Image Processing
Image processing
Chip
Neural Networks
Neural networks
Processing
Data storage equipment
Stream Processing
FPGA Implementation
Unit
Character Recognition
Distributed Memory
Parallel Computation
Operator
Parallel Processing
Data Streams
Character recognition
Artificial Neural Network
Programming

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Kim, D., Kim, H., Kim, H., Han, G., & Chung, D. (2005). A SIMD neural network processor for image processing. Lecture Notes in Computer Science, 3497(II), 665-672.
Kim, Dongsun ; Kim, Hyunsik ; Kim, Hongsik ; Han, Gunhee ; Chung, Duckjin. / A SIMD neural network processor for image processing. In: Lecture Notes in Computer Science. 2005 ; Vol. 3497, No. II. pp. 665-672.
@article{be2a4d7df99247f48f137af0fd7e9124,
title = "A SIMD neural network processor for image processing",
abstract = "Artificial Neural Networks (ANNs) and image processing requires massively parallel computation of simple operator accompanied by heavy memory access. Thus, this type of operators naturally maps onto Single Instruction Multiple Data (SIMD) stream parallel processing with distributed memory. This paper proposes a high performance neural network processor whose function can be changed by programming. The proposed processor is based on the SIMD architecture that is optimized for neural network and image processing. The proposed processor supports 24 instructions, and consists of 16 Processing Units (PUs) per chip. Each PU includes 24-bit 2K-word Local Memory (LM) and a Processing Element (PE). The proposed architecture allows multichip expansion that minimizes chip-to-chip communication bottleneck. The proposed processor is verified with FPGA implementation and the functionality is verified with character recognition application.",
author = "Dongsun Kim and Hyunsik Kim and Hongsik Kim and Gunhee Han and Duckjin Chung",
year = "2005",
month = "9",
day = "26",
language = "English",
volume = "3497",
pages = "665--672",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Springer Verlag",
number = "II",

}

Kim, D, Kim, H, Kim, H, Han, G & Chung, D 2005, 'A SIMD neural network processor for image processing', Lecture Notes in Computer Science, vol. 3497, no. II, pp. 665-672.

A SIMD neural network processor for image processing. / Kim, Dongsun; Kim, Hyunsik; Kim, Hongsik; Han, Gunhee; Chung, Duckjin.

In: Lecture Notes in Computer Science, Vol. 3497, No. II, 26.09.2005, p. 665-672.

Research output: Contribution to journalConference article

TY - JOUR

T1 - A SIMD neural network processor for image processing

AU - Kim, Dongsun

AU - Kim, Hyunsik

AU - Kim, Hongsik

AU - Han, Gunhee

AU - Chung, Duckjin

PY - 2005/9/26

Y1 - 2005/9/26

N2 - Artificial Neural Networks (ANNs) and image processing requires massively parallel computation of simple operator accompanied by heavy memory access. Thus, this type of operators naturally maps onto Single Instruction Multiple Data (SIMD) stream parallel processing with distributed memory. This paper proposes a high performance neural network processor whose function can be changed by programming. The proposed processor is based on the SIMD architecture that is optimized for neural network and image processing. The proposed processor supports 24 instructions, and consists of 16 Processing Units (PUs) per chip. Each PU includes 24-bit 2K-word Local Memory (LM) and a Processing Element (PE). The proposed architecture allows multichip expansion that minimizes chip-to-chip communication bottleneck. The proposed processor is verified with FPGA implementation and the functionality is verified with character recognition application.

AB - Artificial Neural Networks (ANNs) and image processing requires massively parallel computation of simple operator accompanied by heavy memory access. Thus, this type of operators naturally maps onto Single Instruction Multiple Data (SIMD) stream parallel processing with distributed memory. This paper proposes a high performance neural network processor whose function can be changed by programming. The proposed processor is based on the SIMD architecture that is optimized for neural network and image processing. The proposed processor supports 24 instructions, and consists of 16 Processing Units (PUs) per chip. Each PU includes 24-bit 2K-word Local Memory (LM) and a Processing Element (PE). The proposed architecture allows multichip expansion that minimizes chip-to-chip communication bottleneck. The proposed processor is verified with FPGA implementation and the functionality is verified with character recognition application.

UR - http://www.scopus.com/inward/record.url?scp=24944561980&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=24944561980&partnerID=8YFLogxK

M3 - Conference article

VL - 3497

SP - 665

EP - 672

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

IS - II

ER -

Kim D, Kim H, Kim H, Han G, Chung D. A SIMD neural network processor for image processing. Lecture Notes in Computer Science. 2005 Sep 26;3497(II):665-672.