Finding arbitrary shaped clusters for character recognition

Noha Yousri, Mohamed S. Kamel, Mohamed A. Ismail

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

Abstract

Several supervised and unsupervised methods have been applied to the field of character recognition. In this research we focus on the unsupervised methods used to group similar characters together. Instead of using the traditional clustering algorithms, which are mainly restricted to globular-shaped clusters, we use an efficient distance based clustering that identifies the natural shapes of clusters according to their densities. Thus, in the case of character recognition, where it is natural to have different writing styles for the same character, the algorithm can be used to discover the continuity between character feature vectors, which cannot be discovered by traditional algorithms. This paper |introduces the use of an algorithm that efficiently finds arbitrary-shaped clusters of characters, and compares it to related algorithms. Two character recognition data sets are used to illustrate the efficiency of the suggested algorithm.

Original languageEnglish
Title of host publicationImage Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings
Pages597-608
Number of pages12
DOIs
Publication statusPublished - 29 Jul 2008
Externally publishedYes
Event5th International Conference on Image Analysis and Recognition, ICIAR 2008 - Povoa de Varzim, Portugal
Duration: 25 Jun 200827 Jun 2008

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5112 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other5th International Conference on Image Analysis and Recognition, ICIAR 2008
CountryPortugal
CityPovoa de Varzim
Period25/6/0827/6/08

Fingerprint

Character recognition
Character Recognition
Arbitrary
Feature Vector
Clustering algorithms
Clustering Algorithm
Clustering
Character

Keywords

  • Arbitrary shaped clusters
  • Character recognition
  • Clustering

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Yousri, N., Kamel, M. S., & Ismail, M. A. (2008). Finding arbitrary shaped clusters for character recognition. In Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings (pp. 597-608). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5112 LNCS). https://doi.org/10.1007/978-3-540-69812-8_59

Finding arbitrary shaped clusters for character recognition. / Yousri, Noha; Kamel, Mohamed S.; Ismail, Mohamed A.

Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings. 2008. p. 597-608 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5112 LNCS).

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

Yousri, N, Kamel, MS & Ismail, MA 2008, Finding arbitrary shaped clusters for character recognition. in Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5112 LNCS, pp. 597-608, 5th International Conference on Image Analysis and Recognition, ICIAR 2008, Povoa de Varzim, Portugal, 25/6/08. https://doi.org/10.1007/978-3-540-69812-8_59
Yousri N, Kamel MS, Ismail MA. Finding arbitrary shaped clusters for character recognition. In Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings. 2008. p. 597-608. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-540-69812-8_59
Yousri, Noha ; Kamel, Mohamed S. ; Ismail, Mohamed A. / Finding arbitrary shaped clusters for character recognition. Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings. 2008. pp. 597-608 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{684cff0c93494698af16f38b785ba1ce,
title = "Finding arbitrary shaped clusters for character recognition",
abstract = "Several supervised and unsupervised methods have been applied to the field of character recognition. In this research we focus on the unsupervised methods used to group similar characters together. Instead of using the traditional clustering algorithms, which are mainly restricted to globular-shaped clusters, we use an efficient distance based clustering that identifies the natural shapes of clusters according to their densities. Thus, in the case of character recognition, where it is natural to have different writing styles for the same character, the algorithm can be used to discover the continuity between character feature vectors, which cannot be discovered by traditional algorithms. This paper |introduces the use of an algorithm that efficiently finds arbitrary-shaped clusters of characters, and compares it to related algorithms. Two character recognition data sets are used to illustrate the efficiency of the suggested algorithm.",
keywords = "Arbitrary shaped clusters, Character recognition, Clustering",
author = "Noha Yousri and Kamel, {Mohamed S.} and Ismail, {Mohamed A.}",
year = "2008",
month = "7",
day = "29",
doi = "10.1007/978-3-540-69812-8_59",
language = "English",
isbn = "3540698116",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "597--608",
booktitle = "Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings",

}

TY - GEN

T1 - Finding arbitrary shaped clusters for character recognition

AU - Yousri, Noha

AU - Kamel, Mohamed S.

AU - Ismail, Mohamed A.

PY - 2008/7/29

Y1 - 2008/7/29

N2 - Several supervised and unsupervised methods have been applied to the field of character recognition. In this research we focus on the unsupervised methods used to group similar characters together. Instead of using the traditional clustering algorithms, which are mainly restricted to globular-shaped clusters, we use an efficient distance based clustering that identifies the natural shapes of clusters according to their densities. Thus, in the case of character recognition, where it is natural to have different writing styles for the same character, the algorithm can be used to discover the continuity between character feature vectors, which cannot be discovered by traditional algorithms. This paper |introduces the use of an algorithm that efficiently finds arbitrary-shaped clusters of characters, and compares it to related algorithms. Two character recognition data sets are used to illustrate the efficiency of the suggested algorithm.

AB - Several supervised and unsupervised methods have been applied to the field of character recognition. In this research we focus on the unsupervised methods used to group similar characters together. Instead of using the traditional clustering algorithms, which are mainly restricted to globular-shaped clusters, we use an efficient distance based clustering that identifies the natural shapes of clusters according to their densities. Thus, in the case of character recognition, where it is natural to have different writing styles for the same character, the algorithm can be used to discover the continuity between character feature vectors, which cannot be discovered by traditional algorithms. This paper |introduces the use of an algorithm that efficiently finds arbitrary-shaped clusters of characters, and compares it to related algorithms. Two character recognition data sets are used to illustrate the efficiency of the suggested algorithm.

KW - Arbitrary shaped clusters

KW - Character recognition

KW - Clustering

UR - http://www.scopus.com/inward/record.url?scp=47849124761&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47849124761&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-69812-8_59

DO - 10.1007/978-3-540-69812-8_59

M3 - Conference contribution

AN - SCOPUS:47849124761

SN - 3540698116

SN - 9783540698111

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 597

EP - 608

BT - Image Analysis and Recognition - 5th International Conference, ICIAR 2008, Proceedings

ER -