Publication detail

Character-based Writer Verification of Ancient Hebrew Square-script Manuscripts: On Edge-direction Feature

TOBING, T. ŠKRABÁNEK, P. YILDIRIM, S. GEORGE, S. ELGVIN, T.

Original Title

Character-based Writer Verification of Ancient Hebrew Square-script Manuscripts: On Edge-direction Feature

Type

conference paper

Language

English

Original Abstract

Handwriting significantly contributes to the task of the writer identification and verification of modern and historical documents. This work developed a writer verification system for ancient Hebrew square-script manuscripts, mainly based on the edge-direction feature. Two configurations within the proposed system are carried out, i.e., character-based edge-direction feature extraction and extraction techniques of handwriting shape representation that may drive the system performance. A classification-based verification approach, utilizing Support Vector Machine (SVM) as the classifier, is employed to evaluate the performance of the two configurations. This study has confirmed that the skeleton-based shape representation technique outperforms the edge detection technique used in the predecessor approach. Furthermore, a character-based writer verification system provides the corresponding scholars and experts with an alphabetical investigation to identify the uniqueness of each writer’s handwriting.

Keywords

writer verification; edge-direction feature; edge detection; ancient manuscript

Authors

TOBING, T.; ŠKRABÁNEK, P.; YILDIRIM, S.; GEORGE, S.; ELGVIN, T.

Released

23. 6. 2023

ISBN

978-0-89208-362-6

Book

Archiving Conference

Pages from

155

Pages to

158

Pages count

4

URL

BibTex

@inproceedings{BUT185701,
  author="Tabita Lumban {Tobing} and Pavel {Škrabánek} and Sule {Yildirim} and Sony {George} and Torleif {Elgvin}",
  title="Character-based Writer Verification of Ancient Hebrew Square-script Manuscripts: On Edge-direction Feature",
  booktitle="Archiving Conference",
  year="2023",
  pages="155--158",
  doi="10.2352/issn.2168-3204.2023.20.1.32",
  isbn="978-0-89208-362-6",
  url="https://library.imaging.org/archiving/articles/20/1/32"
}