Papperoni

1990

Handwritten Digit Recognition With a Back-Propagation Network

Yann Lecun, B. Boser, John S. Denker, D. Henderson, R. E. Howard, W. Hubbard, L. D. Jackel

citations

Cite Score

AI summary

This paper introduces a convolutional neural network architecture trained with back-propagation for handwritten digit recognition, achieving a 1% error rate on zipcode digits using a dataset of segmented numerals digitized from real U.S. Mail and printed digits.

Main Contributions

Introduces a multi-layer network architecture for handwritten digit recognition.
Demonstrates the application of back-propagation learning to a real-world task with minimal preprocessing.
Achieves a 1% error rate on zipcode digits recognition.
Presents a network design that incorporates geometric knowledge and constraints to improve performance.
Shows the feasibility of implementing the network on commercial digital signal processing hardware for practical applications.

Abstract

We present an application of back-propagation networks to handwritten digit recognition. Minimal preprocessing of the data was required, but architecture of the network was highly constrained and specifically designed for the task. The input of the network consists of normalized images of isolated digits. The method has 1% error rate and about a 9% reject rate on zipcode digits provided by the U.S. Postal Service.

Citation Graph

Loading graph...

References [11]

Sort:

Filter:

[1]Learning Internal Representations by Error Propagation

D. E. Rumelhart, Geoffrey E. Hinton, Ronald J. Williams - 1986

46 papers in library cite

Google Scholar

I expected very little of this, but was so good in explaining concepts! Very good read. It gets a bit boring when it starts explaining things by the end of the chapter, but good nonetheless.

[2]Backpropagation Applied to Handwritten Zip-Code Recognition

Yann Lecun, B. Boser, John S. Denker, D. Henderson, R. E. Howard, W. Hubbard, L. D. Jackal - 1989

24 papers in library cite

Google Scholar

The first convolution NN! Very simple concept and very simply explained. Very good results and overall a good read.

[3]Optimal Brain Damage

Yann Lecun, John Denker, Sara Solla, Richard Howard, Lawrence Jackel - 1990

4 papers in library cite

Google Scholar

It's a nice idea but I think it's a bit uninteresting - maybe it influenced other important work later?

[4]Generalization and Network Design Strategies

Yann Lecun - 1989

5 papers in library cite

Google Scholar

It's a nice intro to LeCun's early work and how he got to the Conv Net

[5]Large Automatic Learning, Rule Extraction and Generalization

John Denker, Daniel Schwartz, Ben Wittner, Sara A. Solla, Richard Howard, Lawrence Jackel, John Hopfield - 1987

4 papers in library cite

Google Scholar

Good paper overall, but things get a bit boring 2/3 of the way through. However, I feel that this is a good piece of early work into NNs.

[6]Neural Network Recognizer for Hand-Written Zip Code Digits

I. Guyon - 1989

2 papers in library cite

Google Scholar

This is probably the best explanation of convolutions I've seen in a paper.

[7]Neocognitron: A New algorithm for Pattern Recognition Tolerant of Deformations and Shifts in Position

Kunihiko Fukushima, S. Miyake - 1982

7 papers in library cite

Google Scholar

[8]Sn: A Simulator for Connectionist Models

L. Y. Bottou, Yann Lecun - 1988

5 papers in library cite

Google Scholar

[9]A Framework for Object Recognition in a Visually Complex Environment and Its Application to Locating Address Blocks on Mail Pieces

C. H. Wang, S. N. Srihari - 1988

3 papers in library cite

Google Scholar

[10]Learning Networks of Neurons With Boolean Logic

S. Patarnello, P. Carnevali - 1987

3 papers in library cite

Google Scholar

[11]Consistent Inference of Probabilities in Layered Networks: Predictions and Generalization

N. Tishby, E. Levin, Sara A. Solla - 1989

2 papers in library cite

Google Scholar

Cited by

papers in your library

Cites

papers in your library

Read

on June 21, 2025

Just a minor derivation of the previous paper, adding more layers to the net. Pretty much the same exact text than the other one. Uninteresting if you read the other.