Position invariant optical character recognition through symmetry features

Sam Holland; Richard Neville

doi:10.1109/ACT.2009.84

Position invariant optical character recognition through symmetry features

Sam Holland, Richard Neville

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

156 Downloads (Pure)

Abstract

We propose an effective method to achieve position invariance in the application of Optical Character Recognition (OCR). We normalise the position of all inputs based on their symmetry features. The Generalized Symmetry Transform (GST) is used to determine the symmetry features prior to classification by a Probabilistic Neural Network (PNN). We used the United States Postal Service (USPS) data set to measure performance. © 2009 IEEE.

Original language	English
Title of host publication	ACT 2009 - International Conference on Advances in Computing, Control and Telecommunication Technologies\|ACT - Int. Conf. Adv. Comput., Control Telecommun. Technol.
Place of Publication	India
Publisher	IEEE
Pages	313-315
Number of pages	2
ISBN (Print)	9780769539157
DOIs	https://doi.org/10.1109/ACT.2009.84
Publication status	Published - 2009
Event	International Conference on Advances in Computing, Control and Telecommunication Technologies, ACT 2009 - Trivandrum, Kerala Duration: 1 Jul 2009 → …

Conference

Conference	International Conference on Advances in Computing, Control and Telecommunication Technologies, ACT 2009
City	Trivandrum, Kerala
Period	1/07/09 → …

Access to Document

10.1109/ACT.2009.84

FULL-TEXT

Cite this

@inproceedings{c5092ead013f46888c3a19785f27b2b9,

title = "Position invariant optical character recognition through symmetry features",

abstract = "We propose an effective method to achieve position invariance in the application of Optical Character Recognition (OCR). We normalise the position of all inputs based on their symmetry features. The Generalized Symmetry Transform (GST) is used to determine the symmetry features prior to classification by a Probabilistic Neural Network (PNN). We used the United States Postal Service (USPS) data set to measure performance. {\textcopyright} 2009 IEEE.",

author = "Sam Holland and Richard Neville",

note = "100 word commentary: A method to achieve position invariance in the application of Optical Character Recognition by normalising the position of all inputs based on their symmetry features; utilises the Generalized Symmetry Transform to determine the symmetry features prior to classification by a Probabilistic Neural Network. The results showed that an affine transformation centring a pattern on the point with the highest symmetry magnitude was effective as a position invariant mechanism for classification. Role: principal academic supervisor of first author who was a PhD student - contribution estimate = 35%. Conference is one of the better Computation Conferences. Most likely REF score 2*.; International Conference on Advances in Computing, Control and Telecommunication Technologies, ACT 2009 ; Conference date: 01-07-2009",

year = "2009",

doi = "10.1109/ACT.2009.84",

language = "English",

isbn = "9780769539157",

pages = "313--315",

booktitle = "ACT 2009 - International Conference on Advances in Computing, Control and Telecommunication Technologies|ACT - Int. Conf. Adv. Comput., Control Telecommun. Technol.",

publisher = "IEEE",

address = "United States",

}

Holland, S & Neville, R 2009, Position invariant optical character recognition through symmetry features. in ACT 2009 - International Conference on Advances in Computing, Control and Telecommunication Technologies|ACT - Int. Conf. Adv. Comput., Control Telecommun. Technol.. IEEE, India, pp. 313-315, International Conference on Advances in Computing, Control and Telecommunication Technologies, ACT 2009, Trivandrum, Kerala, 1/07/09. https://doi.org/10.1109/ACT.2009.84

Position invariant optical character recognition through symmetry features. / Holland, Sam; Neville, Richard.

ACT 2009 - International Conference on Advances in Computing, Control and Telecommunication Technologies|ACT - Int. Conf. Adv. Comput., Control Telecommun. Technol.. India : IEEE, 2009. p. 313-315.

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

TY - GEN

T1 - Position invariant optical character recognition through symmetry features

AU - Holland, Sam

AU - Neville, Richard

N1 - 100 word commentary: A method to achieve position invariance in the application of Optical Character Recognition by normalising the position of all inputs based on their symmetry features; utilises the Generalized Symmetry Transform to determine the symmetry features prior to classification by a Probabilistic Neural Network. The results showed that an affine transformation centring a pattern on the point with the highest symmetry magnitude was effective as a position invariant mechanism for classification. Role: principal academic supervisor of first author who was a PhD student - contribution estimate = 35%. Conference is one of the better Computation Conferences. Most likely REF score 2*.

PY - 2009

Y1 - 2009

N2 - We propose an effective method to achieve position invariance in the application of Optical Character Recognition (OCR). We normalise the position of all inputs based on their symmetry features. The Generalized Symmetry Transform (GST) is used to determine the symmetry features prior to classification by a Probabilistic Neural Network (PNN). We used the United States Postal Service (USPS) data set to measure performance. © 2009 IEEE.

AB - We propose an effective method to achieve position invariance in the application of Optical Character Recognition (OCR). We normalise the position of all inputs based on their symmetry features. The Generalized Symmetry Transform (GST) is used to determine the symmetry features prior to classification by a Probabilistic Neural Network (PNN). We used the United States Postal Service (USPS) data set to measure performance. © 2009 IEEE.

U2 - 10.1109/ACT.2009.84

DO - 10.1109/ACT.2009.84

M3 - Conference contribution

SN - 9780769539157

SP - 313

EP - 315

BT - ACT 2009 - International Conference on Advances in Computing, Control and Telecommunication Technologies|ACT - Int. Conf. Adv. Comput., Control Telecommun. Technol.

PB - IEEE

CY - India

T2 - International Conference on Advances in Computing, Control and Telecommunication Technologies, ACT 2009

Y2 - 1 July 2009

ER -

Position invariant optical character recognition through symmetry features

Abstract

Conference

Access to Document

Fingerprint

Cite this