Back to search

National and Kapodistrian University of Athens

Αναγνώριση χαρακτήρων (OCR) ιστορικών εγγράφων: χειρόγραφα, καλλιγραφικά και τυπωμένα έγγραφα

Abstract

dc:description

“Character recognition” refers to the procedure of ‘reading’ text using a computer, taking as input a document image as well as to the conversion of the document image to electronic text. This dissertation focuses on the segmentation of handwritten document images to the basic semantic units that comprise them, namely text lines and words. Concerning the problem of text line segmentation, we developed a new methodology whose novelties are: (i) an efficient block-based Hough transform in which voting occurs on the basis of equally spaced blocks after splitting of the connected components’ bounding box; (ii) a partitioning of the connected component domain into three spatial sub-domains, for which a different processing strategy of the corresponding connected components can be employed; and (iii) the efficient separation of vertically connected parts of text lines. The proposed text line segmentation methodology has been evaluated against other state-of-the-art text line segmentation methodologies and has proven to achieve better results. Concerning the word segmentation stage, we developed two different methodologies. Concerning the first methodology, the decision on whether a gap is between two words or inside a single word, we proposed a threshold which is calculated making use of several characteristics of the document image. On the second approach, we make use of a well-known methodology in the field of unsupervised clustering, the Gaussian mixture modeling in order to classify the gaps into each class. Experimental results prove the efficiency of the proposed methodologies. Finally, a novel two stage evaluation methodology for word segmentation techniques is proposed. This methodology treats the distance computation stage and the gap classification stage independently, in contrast to current evaluation methodologies for word segmentation

Degree

thesis:*
Grantor dc:publisher
National and Kapodistrian University of Athens
Year dc:date
2009

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Louloudis, Georgios
  • Λουλούδης, Γεώργιος

Subjects

dc:subject × 14

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/23752

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Louloudis, Georgios; Λουλούδης, Γεώργιος. Αναγνώριση χαρακτήρων (OCR) ιστορικών εγγράφων: χειρόγραφα, καλλιγραφικά και τυπωμένα έγγραφα. National and Kapodistrian University of Athens, 2009. http://hdl.handle.net/10442/hedi/23752