Document Examiner Feature Extraction: Thinned vs. Skeletonised Handwriting Images

Vladimir Pervouchine; Graham Leedham

doi:10.1109/TENCON.2005.301018

Document Examiner Feature Extraction: Thinned vs. Skeletonised Handwriting Images

Vladimir Pervouchine
, Graham Leedham

Computer Science

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

5 Citations (Scopus)

Abstract

This paper describes two approaches to approximation of handwriting strokes for use in writer identification. One approach is based on a thinning method and produces raster skeleton whereas the other approximates handwriting strokes by cubic splines and produces a vector skeleton. The vector skeletonisation method is designed to preserve the individual features that can distinguish one writer from another. Extraction of structural character-level features of handwriting is performed using both skeletonisation methods and the results are compared. Use of the vector skeletonisation method resulted in lower error rate during the feature extraction stage. It also enabled to extract more structural features and improved the accuracy of writer identification from 78% to 98% in the experiment with 100 samples of grapheme "th" collected from 20 writers.

Original language	English
Title of host publication	Proceedings of the IEEE Region 10 Conference (TENCON'05)
Place of Publication	Los Alamitos, United States of America
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	2338-2343
ISBN (Print)	0780393112
DOIs	https://doi.org/10.1109/TENCON.2005.301018
Publication status	Published - 2005
Event	TENCON 2005: IEEE Region 10 Conference - Melbourne, Australia Duration: 21 Nov 2005 → 24 Nov 2005

Conference

Conference	TENCON 2005: IEEE Region 10 Conference
City	Melbourne, Australia
Period	21/11/05 → 24/11/05

Keywords

Pattern Recognition and Data Mining
Artificial Intelligence and Image Processing
Image Processing

Access to Document

10.1109/TENCON.2005.301018

https://hdl.handle.net/1959.11/6622

Handle.net

Persistent link

Cite this

@inproceedings{cb6ceea1c16a4b7fb9a709c42c95b7ba,

title = "Document Examiner Feature Extraction: Thinned vs. Skeletonised Handwriting Images",

abstract = "This paper describes two approaches to approximation of handwriting strokes for use in writer identification. One approach is based on a thinning method and produces raster skeleton whereas the other approximates handwriting strokes by cubic splines and produces a vector skeleton. The vector skeletonisation method is designed to preserve the individual features that can distinguish one writer from another. Extraction of structural character-level features of handwriting is performed using both skeletonisation methods and the results are compared. Use of the vector skeletonisation method resulted in lower error rate during the feature extraction stage. It also enabled to extract more structural features and improved the accuracy of writer identification from 78\% to 98\% in the experiment with 100 samples of grapheme {"}th{"} collected from 20 writers.",

keywords = "Pattern Recognition and Data Mining, Artificial Intelligence and Image Processing, Image Processing",

author = "Vladimir Pervouchine and Graham Leedham",

year = "2005",

doi = "10.1109/TENCON.2005.301018",

language = "English",

isbn = "0780393112",

pages = "2338--2343",

booktitle = "Proceedings of the IEEE Region 10 Conference (TENCON'05)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

note = "TENCON 2005: IEEE Region 10 Conference ; Conference date: 21-11-2005 Through 24-11-2005",

}

Pervouchine, V & Leedham, G 2005, Document Examiner Feature Extraction: Thinned vs. Skeletonised Handwriting Images. in Proceedings of the IEEE Region 10 Conference (TENCON'05). Institute of Electrical and Electronics Engineers (IEEE), Los Alamitos, United States of America, pp. 2338-2343, TENCON 2005: IEEE Region 10 Conference, Melbourne, Australia, 21/11/05. https://doi.org/10.1109/TENCON.2005.301018

Document Examiner Feature Extraction: Thinned vs. Skeletonised Handwriting Images. / Pervouchine, Vladimir; Leedham, Graham.
Proceedings of the IEEE Region 10 Conference (TENCON'05). Los Alamitos, United States of America: Institute of Electrical and Electronics Engineers (IEEE), 2005. p. 2338-2343.

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

TY - GEN

T1 - Document Examiner Feature Extraction

T2 - TENCON 2005: IEEE Region 10 Conference

AU - Pervouchine, Vladimir

AU - Leedham, Graham

PY - 2005

Y1 - 2005

N2 - This paper describes two approaches to approximation of handwriting strokes for use in writer identification. One approach is based on a thinning method and produces raster skeleton whereas the other approximates handwriting strokes by cubic splines and produces a vector skeleton. The vector skeletonisation method is designed to preserve the individual features that can distinguish one writer from another. Extraction of structural character-level features of handwriting is performed using both skeletonisation methods and the results are compared. Use of the vector skeletonisation method resulted in lower error rate during the feature extraction stage. It also enabled to extract more structural features and improved the accuracy of writer identification from 78% to 98% in the experiment with 100 samples of grapheme "th" collected from 20 writers.

AB - This paper describes two approaches to approximation of handwriting strokes for use in writer identification. One approach is based on a thinning method and produces raster skeleton whereas the other approximates handwriting strokes by cubic splines and produces a vector skeleton. The vector skeletonisation method is designed to preserve the individual features that can distinguish one writer from another. Extraction of structural character-level features of handwriting is performed using both skeletonisation methods and the results are compared. Use of the vector skeletonisation method resulted in lower error rate during the feature extraction stage. It also enabled to extract more structural features and improved the accuracy of writer identification from 78% to 98% in the experiment with 100 samples of grapheme "th" collected from 20 writers.

KW - Pattern Recognition and Data Mining

KW - Artificial Intelligence and Image Processing

KW - Image Processing

UR - http://www3.ntu.edu.sg/SCE/labs/forse/PDF/docExam_12.pdf

U2 - 10.1109/TENCON.2005.301018

DO - 10.1109/TENCON.2005.301018

M3 - Conference contribution

SN - 0780393112

SP - 2338

EP - 2343

BT - Proceedings of the IEEE Region 10 Conference (TENCON'05)

PB - Institute of Electrical and Electronics Engineers (IEEE)

CY - Los Alamitos, United States of America

Y2 - 21 November 2005 through 24 November 2005

ER -

Document Examiner Feature Extraction: Thinned vs. Skeletonised Handwriting Images

Abstract

Conference

Keywords

Access to Document

Handle.net

Fingerprint

Cite this