Handwritten Digit Recognition with a Back-Propagation Network

Yann LeCun; Bernhard E. Boser; John S. Denker; D. Henderson; Richard Howard; Wayne E. Hubbard; Lawrence D. Jackel

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.

Keywords

Numerical digitComputer sciencePreprocessorWord error rateTask (project management)Speech recognitionArtificial intelligencePostal servicePattern recognition (psychology)ArithmeticMathematicsEngineering

Affiliated Institutions

Related Publications

Improving Performance in Neural Networks Using a Boosting Algorithm

Harris Drucker , Robert E. Schapire , Patrice Simard

A boosting algorithm converts a learning machine with error rate less than 50% to one with an arbitrarily low error rate. However, the algorithm discussed here depends on having...

1992 Neural Information Processing Systems 158 citations

BOOSTING PERFORMANCE IN NEURAL NETWORKS

Harris Drucker , Robert E. Schapire , Patrice Simard

A boosting algorithm, based on the probably approximately correct (PAC) learning model is used to construct an ensemble of neural networks that significantly improves performanc...

1993 International Journal of Pattern Reco... 184 citations

Gradient-based learning applied to document recognition

Yann LeCun , Léon Bottou , Yoshua Bengio +1 more

Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate networ...

1998 Proceedings of the IEEE 55110 citations

A Novel Connectionist System for Unconstrained Handwriting Recognition

Alexander Graves , Marcus Liwicki , S. George Fernandez +3 more

Recognizing lines of unconstrained handwritten text is a challenging task. The difficulty of segmenting cursive or overlapping characters, combined with the need to exploit surr...

2009 IEEE Transactions on Pattern Analysis... 1961 citations

Lattice-based optimization of sequence classification criteria for neural-network acoustic modeling

Brian Kingsbury

Acoustic models used in hidden Markov model/neural-network (HMM/NN) speech recognition systems are usually trained with a frame-based cross-entropy error criterion. In contrast,...

2009 238 citations

Publication Info

Year: 1989
Type: article
Volume: 2
Pages: 396-404
Citations: 3629
Access: Closed

External Links

Citation Metrics

3629

OpenAlex

Cite This

APA Style

                            
                                    Yann LeCun, 
                                
                                    Bernhard E. Boser, 
                                
                                    John S. Denker
                                
                                et al.
                            
                            (1989). 
                            Handwritten Digit Recognition with a Back-Propagation Network. 
                            neural information processing systems
                            , 2
                            
                            , 396-404.