A SIMD neural network processor for image processing

Authors:
Dongsun Kim;Hyunsik Kim;Hongsik Kim;Gunhee Han;Duckjin Chung
Affiliations:
Korea Electronics Technology Institute, PyungTaek-Si, KyungGi-Do, Korea;Korea Electronics Technology Institute, PyungTaek-Si, KyungGi-Do, Korea;Department of Electrical & Electronic Engineering, Yousei University, Seoul, Korea;Department of Electrical & Electronic Engineering, Yousei University, Seoul, Korea;Information Technology and Telecommunications, Inha University, Nam-Gu, Incheon, Korea
Venue:
ISNN'05 Proceedings of the Second international conference on Advances in neural networks - Volume Part II
Year:
2005

Citing 4
Cited 3

Active pages: a computation model for intelligent memory

Proceedings of the 25th annual international symposium on Computer architecture
Pipelined and Parallel Computer Architecture

Pipelined and Parallel Computer Architecture
FlexRAM: Toward an Advanced Intelligent Memory System

ICCD '99 Proceedings of the 1999 IEEE International Conference on Computer Design
Morphological operators for image and video compression

IEEE Transactions on Image Processing

Design and Implementation of a General Purpose Neural Network Processor

ISNN '07 Proceedings of the 4th international symposium on Neural Networks: Part II--Advances in Neural Networks
Artificial neural networks in hardware: A survey of two decades of progress

Neurocomputing
Implementation of a neural network processor based on RISC architecture for various signal processing applications

ISNN'06 Proceedings of the Third international conference on Advances in Neural Networks - Volume Part III

Quantified Score

Hi-index	0.00

Visualization

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.