Chapter

Advances in Neural Networks – ISNN 2005

Volume 3497 of the series Lecture Notes in Computer Science pp 665-672

A SIMD Neural Network Processor for Image Processing

  • Dongsun KimAffiliated withLancaster UniversityDMB Project Office, Korea Electronics Technology Institute
  • , Hyunsik KimAffiliated withLancaster UniversityDMB Project Office, Korea Electronics Technology Institute
  • , Hongsik KimAffiliated withCarnegie Mellon UniversityDepartment of Electrical & Electronic Engineering, Yousei University
  • , Gunhee HanAffiliated withCarnegie Mellon UniversityDepartment of Electrical & Electronic Engineering, Yousei University
  • , Duckjin ChungAffiliated withCarnegie Mellon UniversityInformation Technology and Telecommunications, Inha University

* Final gross prices may vary according to local VAT.

Get Access

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 multi-chip 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.