Text line extraction and baseline detection
Pages 194 - 210
Abstract
A hypothesis-driven analysis of text regions in document images is presented. We assume that document images are constructed by the TeX imaging model (the boxes and glue model), and introduce the box representation of a printed symbol called character prototype. The algorithm estimates the location of an hbox (line of text) in a vbox (column of text). This hypothesis is accepted if most connected components in the hbox form character boxes. Character prototypes are used to determine whether a box in an hbox is a character box. The algorithm correctly extracted an average of 96--99 percent of text-lines from digitized text-columns written in English. The character prototype scheme also accurately represented the Japanese alphabet, Chinese characters, and Bengali words, and text-lines were correctly extracted from documents written in these languages.
References
[1]
{AKM71} R. Ascher, G. Koppelman, M. Miller, G. Nagy, and G. Shelton, Jr., "An Interactive System for Reading Unformatted Printed Text," IEEE Transactions on Computers, Vol. C-20, No. 12, pp. 1527--1543, December 1971.
[2]
{Bai88} H. Baird, "Global-to-Local Layout Analysis," IAPR Proceedings of the Workshop on Syntactic and Structural Pattern Recognition, Pont-a-Mousson, France, pp. 136--146 September 1988.
[3]
{HH85} M. Hase, and Y. Hoshino, "Segmentation Method of Document Images by Two-Dimensional Fourier Transformation," System and Computers in Japan, Vol. 16, No. 3, pp. 38--47, 1985.
[4]
{Kan90} J. Kanai, "Text-Line Extraction Using Character Prototypes," Pre-Proceedings of International Association for Pattern Recognition Workshop on Syntactic & Structural Pattern Recognition, Murray Hill, New Jersey. June 13--15, 1990, pp. 182--191.
[5]
{Kat86} T. Kato, Theoretical and Experimental Study on Parallel Processing in Real-Time Layout Understanding and Article Extraction of Document Images, Doctorat Thesis, Kyoto University, Nov. 1986.
[6]
{Knu84} D. Knuth, The TeXbook, Addison-Wesley, 1984.
[7]
{Knu86} D. Knuth, Computers & Typesetting, Vol. E: Computer Modern Typefaces, Addison-Wesley, 1986.
[8]
{NKK88} G. Nagy, J. Kanai, M. Krishnamoorthy, M. Thomas, and M. Viswanathan, "Two Complementary Techniques for Digitized Document Analysis," Proceedings: ACM Conference on Document Processing Systems, Santa Fe, New Mexico, pp. 169--176, December 1988.
[9]
{NS84} G. Nagy, and S. Seth, "Hierarchical Representation of Optically Scanned Documents," Proceedings of the 7th International Conference on Pattern Recognition, Montreal, Canada, pp. 347--349, 1984.
[10]
{NSS86} G. Nagy, S. Seth, and S. Stoddard, "Document Analysis with an Expert System," Pattern Recognition in Practice II, ed. E. S. Gelsema and L. N. Kanal, Elsevier Science Publishers B. V., North-Holland, pp. 149--159, 1986.
[11]
{OM90} M. Okamoto and A. Miyazawa, "An Experimental Implementation of Document Recognition System for Papers Containing Mathematical Expressions," Pre-Proceedings of International Association for Pattern Recognition Workshop on Syntactic & Structural Pattern Recognition, Murray Hill, New Jersey, June 13--15, 1990, pp. 335--350.
[12]
{Ots79} N. Otsu, "A Threshold Selection Method from Gray-Level Histograms," IEEE SMC, Vol. SMC-9, No. 1, January 1979, pp. 62--66.
Recommendations
Towards Generic Text-Line Extraction
ICDAR '13: Proceedings of the 2013 12th International Conference on Document Analysis and RecognitionText-line extraction is the backbone of document image analysis. Since decades, a large number of text-line finding methods have been proposed, where these methods rely on certain assumptions about a target class of documents with respect to writing ...
Baseline detection of multi-lingual unconstrained handwritten text lines
A novel method for baseline detection of multi-lingual multi-oriented text lines.To our knowledge, this is the first baseline detection method for multi-turn text lines.The method uses machine learning along with rotation invariant features for ...
Distance transform based text-line extraction from unconstrained handwritten document images
Highlights- Uses mostly dynamic parameters compatible with any type of handwritten document.
- Deals with the non-uniform gaps between consecutive text-lines and words.
- Generates separation seams based on heuristics with empirically computed ...
AbstractText-line extraction (TLE) is the process of segmenting a document page into lines of text for processing by modules such as language and writer identification or Optical Character Recognition (OCR). Designing of an appropriate TLE method is ...
Comments
Information & Contributors
Information
Published In
Publisher
LE CENTRE DE HAUTES ETUDES INTERNATIONALES D'INFORMATIQUE DOCUMENTAIRE
Paris, France
Publication History
Published: 02 April 1991
Qualifiers
- Research-article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 362Total Downloads
- Downloads (Last 12 months)50
- Downloads (Last 6 weeks)11
Reflects downloads up to 01 Nov 2024
Other Metrics
Citations
View Options
Get Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in