×
Copy the page URI to the clipboard
Säckinger, Eduard; Boser, Bernhard E.; Bromley, Jane M.; LeCun, Yann and Jackel, Larry D.
(1992).
DOI: https://doi.org/10.1109/72.129422
Abstract
A neural network with 136000 connections for recognition of handwritten digits has been implemented using a mixed analog/digital neural network chip. The neural network chip is capable of processing 1000 characters/s. The recognition system has essentially the same rate (5%) as a simulation of the network with 32-b floating-point precision