Application of the ANNA Neural Network Chip to High-Speed Character Recognition

Eduard Sackinger, Bernhard E. Boser, Jane Bromley, Yann LeCun, Lawrence D. Jackel

Research output: Contribution to journalArticle

Abstract

A neural network with 136000 connections for recognition of handwritten digits has been implemented using a mixed analog/digital neural network ship. The neural network chip is capable of processing 1000 characters per second. The recognition system has essentially the same error rate (5%) as a simulation of the network with 32-bit floating-point precision.

Original languageEnglish (US)
Pages (from-to)498-505
Number of pages8
JournalIEEE Transactions on Neural Networks
Volume3
Issue number3
DOIs
StatePublished - May 1992

    Fingerprint

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this