Massive character recognition with a large ground-truthed database

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In character recognition, multiple prototype classifiers, where multiple patterns are prepared as representative patterns of each class, have often been employed to improve recognition accuracy. Our question is how we can improve the recognition accuracy by increasing prototypes massively in the multiple prototype classifier. In this paper, we will answer this question through several experimental analyses, using a simple 1-nearest neighbor (1-NN) classifier and about 550,000 manually labeled handwritten numeral patterns. The analysis results under the leave-one-out evaluation showed not only a simple fact that more prototypes provide fewer recognition errors, but also a more important fact that the error rate decreases approximately to 40% by increasing the prototypes 10 times. The analysis results also showed other phenomena in massive character recognition, such that the NN prototypes become visually closer to the input pattern by increasing the prototypes.

Original languageEnglish
Title of host publication26th Annual ACM Symposium on Applied Computing, SAC 2011
Pages240-244
Number of pages5
DOIs
Publication statusPublished - Jun 23 2011
Event26th Annual ACM Symposium on Applied Computing, SAC 2011 - TaiChung, Taiwan, Province of China
Duration: Mar 21 2011Mar 24 2011

Publication series

NameProceedings of the ACM Symposium on Applied Computing

Other

Other26th Annual ACM Symposium on Applied Computing, SAC 2011
CountryTaiwan, Province of China
CityTaiChung
Period3/21/113/24/11

Fingerprint

Character recognition
Classifiers

All Science Journal Classification (ASJC) codes

  • Software

Cite this

Cai, W., Feng, Y., & Uchida, S. (2011). Massive character recognition with a large ground-truthed database. In 26th Annual ACM Symposium on Applied Computing, SAC 2011 (pp. 240-244). (Proceedings of the ACM Symposium on Applied Computing). https://doi.org/10.1145/1982185.1982241

Massive character recognition with a large ground-truthed database. / Cai, Wenjie; Feng, Yaokai; Uchida, Seiichi.

26th Annual ACM Symposium on Applied Computing, SAC 2011. 2011. p. 240-244 (Proceedings of the ACM Symposium on Applied Computing).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Cai, W, Feng, Y & Uchida, S 2011, Massive character recognition with a large ground-truthed database. in 26th Annual ACM Symposium on Applied Computing, SAC 2011. Proceedings of the ACM Symposium on Applied Computing, pp. 240-244, 26th Annual ACM Symposium on Applied Computing, SAC 2011, TaiChung, Taiwan, Province of China, 3/21/11. https://doi.org/10.1145/1982185.1982241
Cai W, Feng Y, Uchida S. Massive character recognition with a large ground-truthed database. In 26th Annual ACM Symposium on Applied Computing, SAC 2011. 2011. p. 240-244. (Proceedings of the ACM Symposium on Applied Computing). https://doi.org/10.1145/1982185.1982241
Cai, Wenjie ; Feng, Yaokai ; Uchida, Seiichi. / Massive character recognition with a large ground-truthed database. 26th Annual ACM Symposium on Applied Computing, SAC 2011. 2011. pp. 240-244 (Proceedings of the ACM Symposium on Applied Computing).
@inproceedings{5a55ee466d724571a1151009d19d2d73,
title = "Massive character recognition with a large ground-truthed database",
abstract = "In character recognition, multiple prototype classifiers, where multiple patterns are prepared as representative patterns of each class, have often been employed to improve recognition accuracy. Our question is how we can improve the recognition accuracy by increasing prototypes massively in the multiple prototype classifier. In this paper, we will answer this question through several experimental analyses, using a simple 1-nearest neighbor (1-NN) classifier and about 550,000 manually labeled handwritten numeral patterns. The analysis results under the leave-one-out evaluation showed not only a simple fact that more prototypes provide fewer recognition errors, but also a more important fact that the error rate decreases approximately to 40{\%} by increasing the prototypes 10 times. The analysis results also showed other phenomena in massive character recognition, such that the NN prototypes become visually closer to the input pattern by increasing the prototypes.",
author = "Wenjie Cai and Yaokai Feng and Seiichi Uchida",
year = "2011",
month = "6",
day = "23",
doi = "10.1145/1982185.1982241",
language = "English",
isbn = "9781450301138",
series = "Proceedings of the ACM Symposium on Applied Computing",
pages = "240--244",
booktitle = "26th Annual ACM Symposium on Applied Computing, SAC 2011",

}

TY - GEN

T1 - Massive character recognition with a large ground-truthed database

AU - Cai, Wenjie

AU - Feng, Yaokai

AU - Uchida, Seiichi

PY - 2011/6/23

Y1 - 2011/6/23

N2 - In character recognition, multiple prototype classifiers, where multiple patterns are prepared as representative patterns of each class, have often been employed to improve recognition accuracy. Our question is how we can improve the recognition accuracy by increasing prototypes massively in the multiple prototype classifier. In this paper, we will answer this question through several experimental analyses, using a simple 1-nearest neighbor (1-NN) classifier and about 550,000 manually labeled handwritten numeral patterns. The analysis results under the leave-one-out evaluation showed not only a simple fact that more prototypes provide fewer recognition errors, but also a more important fact that the error rate decreases approximately to 40% by increasing the prototypes 10 times. The analysis results also showed other phenomena in massive character recognition, such that the NN prototypes become visually closer to the input pattern by increasing the prototypes.

AB - In character recognition, multiple prototype classifiers, where multiple patterns are prepared as representative patterns of each class, have often been employed to improve recognition accuracy. Our question is how we can improve the recognition accuracy by increasing prototypes massively in the multiple prototype classifier. In this paper, we will answer this question through several experimental analyses, using a simple 1-nearest neighbor (1-NN) classifier and about 550,000 manually labeled handwritten numeral patterns. The analysis results under the leave-one-out evaluation showed not only a simple fact that more prototypes provide fewer recognition errors, but also a more important fact that the error rate decreases approximately to 40% by increasing the prototypes 10 times. The analysis results also showed other phenomena in massive character recognition, such that the NN prototypes become visually closer to the input pattern by increasing the prototypes.

UR - http://www.scopus.com/inward/record.url?scp=79959304421&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79959304421&partnerID=8YFLogxK

U2 - 10.1145/1982185.1982241

DO - 10.1145/1982185.1982241

M3 - Conference contribution

AN - SCOPUS:79959304421

SN - 9781450301138

T3 - Proceedings of the ACM Symposium on Applied Computing

SP - 240

EP - 244

BT - 26th Annual ACM Symposium on Applied Computing, SAC 2011

ER -