Ensemble methods for handwritten digit recognition

Lars Kai Hansen; Christian Liisberg; P. Salamon

doi:10.1109/NNSP.1992.253679

Ensemble methods for handwritten digit recognition

Lars Kai Hansen
, Christian Liisberg
, P. Salamon

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

907 Downloads (Orbit)

Abstract

Neural network ensembles are applied to handwritten digit recognition. The individual networks of the ensemble are combinations of sparse look-up tables (LUTs) with random receptive fields. It is shown that the consensus of a group of networks outperforms the best individual of the ensemble. It is further shown that it is possible to estimate the ensemble performance as well as the learning curve on a medium-size database. In addition the authors present preliminary analysis of experiments on a large database and show that state-of-the-art performance can be obtained using the ensemble approach by optimizing the receptive fields. It is concluded that it is possible to improve performance significantly by introducing moderate-size ensembles; in particular, a 20-25% improvement has been found. The ensemble random LUTs, when trained on a medium-size database, reach a performance (without rejects) of 94% correct classification on digits written by an independent group of people

Original language	English
Title of host publication	Proceedings of the IEEE-SP Workshop Neural Networks for Signal Processing
Publisher	IEEE
Publication date	1992
ISBN (Print)	0-7803-0557-4
DOIs	https://doi.org/10.1109/NNSP.1992.253679
Publication status	Published - 1992
Event	1992 IEEE Workshop on Neural Networks for Signal Processing - Hotel Marielyst, Helsingoer, Denmark Duration: 31 Aug 1992 → 2 Sept 1992 http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=631

Conference

Conference	1992 IEEE Workshop on Neural Networks for Signal Processing
Location	Hotel Marielyst
Country/Territory	Denmark
City	Helsingoer
Period	31/08/1992 → 02/09/1992
Internet address	http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=631

Bibliographical note

Copyright: 1992 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE

Access to Document

10.1109/NNSP.1992.253679

LiisbergFinal published version, 487 KB

OpenUrl availability

Check availability in DTU Findit

Cite this

@inproceedings{400b059284af408ea1af58c120856056,

title = "Ensemble methods for handwritten digit recognition",

abstract = "Neural network ensembles are applied to handwritten digit recognition. The individual networks of the ensemble are combinations of sparse look-up tables (LUTs) with random receptive fields. It is shown that the consensus of a group of networks outperforms the best individual of the ensemble. It is further shown that it is possible to estimate the ensemble performance as well as the learning curve on a medium-size database. In addition the authors present preliminary analysis of experiments on a large database and show that state-of-the-art performance can be obtained using the ensemble approach by optimizing the receptive fields. It is concluded that it is possible to improve performance significantly by introducing moderate-size ensembles; in particular, a 20-25\% improvement has been found. The ensemble random LUTs, when trained on a medium-size database, reach a performance (without rejects) of 94\% correct classification on digits written by an independent group of people",

author = "Hansen, \{Lars Kai\} and Christian Liisberg and P. Salamon",

note = "Copyright: 1992 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE; 1992 IEEE Workshop on Neural Networks for Signal Processing, IEEE NNSP 1992 ; Conference date: 31-08-1992 Through 02-09-1992",

year = "1992",

doi = "10.1109/NNSP.1992.253679",

language = "English",

isbn = "0-7803-0557-4",

booktitle = "Proceedings of the IEEE-SP Workshop Neural Networks for Signal Processing",

publisher = "IEEE",

address = "United States",

url = "http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=631",

}

TY - GEN

T1 - Ensemble methods for handwritten digit recognition

AU - Hansen, Lars Kai

AU - Liisberg, Christian

AU - Salamon, P.

N1 - Copyright: 1992 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE

PY - 1992

Y1 - 1992

N2 - Neural network ensembles are applied to handwritten digit recognition. The individual networks of the ensemble are combinations of sparse look-up tables (LUTs) with random receptive fields. It is shown that the consensus of a group of networks outperforms the best individual of the ensemble. It is further shown that it is possible to estimate the ensemble performance as well as the learning curve on a medium-size database. In addition the authors present preliminary analysis of experiments on a large database and show that state-of-the-art performance can be obtained using the ensemble approach by optimizing the receptive fields. It is concluded that it is possible to improve performance significantly by introducing moderate-size ensembles; in particular, a 20-25% improvement has been found. The ensemble random LUTs, when trained on a medium-size database, reach a performance (without rejects) of 94% correct classification on digits written by an independent group of people

AB - Neural network ensembles are applied to handwritten digit recognition. The individual networks of the ensemble are combinations of sparse look-up tables (LUTs) with random receptive fields. It is shown that the consensus of a group of networks outperforms the best individual of the ensemble. It is further shown that it is possible to estimate the ensemble performance as well as the learning curve on a medium-size database. In addition the authors present preliminary analysis of experiments on a large database and show that state-of-the-art performance can be obtained using the ensemble approach by optimizing the receptive fields. It is concluded that it is possible to improve performance significantly by introducing moderate-size ensembles; in particular, a 20-25% improvement has been found. The ensemble random LUTs, when trained on a medium-size database, reach a performance (without rejects) of 94% correct classification on digits written by an independent group of people

U2 - 10.1109/NNSP.1992.253679

DO - 10.1109/NNSP.1992.253679

M3 - Article in proceedings

SN - 0-7803-0557-4

BT - Proceedings of the IEEE-SP Workshop Neural Networks for Signal Processing

PB - IEEE

T2 - 1992 IEEE Workshop on Neural Networks for Signal Processing

Y2 - 31 August 1992 through 2 September 1992

ER -

Ensemble methods for handwritten digit recognition

Abstract

Conference

Bibliographical note

Access to Document

OpenUrl availability

Fingerprint

Cite this