Elamur Research⌕
☰
home / papers / 2021 / Off-Line Arabic Handwritten Words Segmentation using Morphological Operators

Off-Line Arabic Handwritten Words Segmentation using Morphological Operators

Summary

The paper addresses segmentation of offline handwritten Arabic words into sub-words, also called pieces of Arabic words (PAWs). Arabic is cursive, and six letters never join to the following letter, so a word naturally breaks into several connected pieces. Segmentation becomes unreliable when strokes are accidentally broken by pen lifts, faded ink, scanning, or binarization, and the authors identify such gaps as a major source of errors.

They propose a three-stage pipeline:

  • Pre-processing: binarization, a combination of three morphological operators to connect gaps (the "CGs" step), and skeletonization.
  • Segmentation: small diacritics such as dots are removed, then each remaining connected component is enclosed in a bounding box as a sub-word.
  • Evaluation: results are scored automatically against database ground truth.

The authors chose IESK-ArDB because, unlike IFN/ENIT, it provides sub-word positions in its ground truth. On 1,131 randomly selected images, the method reached 88% agreement with the ground truth after gap connection, which the authors report as higher than comparable earlier work.

Key points

  • Arabic has 28 letters, each with 2–5 positional shapes. Fifteen letters carry dots, and six connect only on one side, which produces the sub-word (PAW) structure.
  • The evaluation used 1,131 randomly chosen images from IESK-ArDB, and segmentation accuracy reached 88% against the database ground truth.
  • The method uses morphological gap closing in pre-processing, then diacritic removal and connected-component bounding boxes for segmentation.
  • Comparison points from related work:
    • AlKhateeb et al. reached 85% on 200 IFN/ENIT images.
    • Ghaleb et al.'s pushing technique reached 72%.
    • A bounding-box distance method tested on 450 IESK-ArDB images reported no overall accuracy.
  • IFN/ENIT lacks character- or sub-word-level positions, so segmentation scored on it needs manually added annotation files.

Interesting ideas

  • I1. Repair accidental fragmentation before splitting into units, so that segmentation boundaries reflect real structure rather than noise. — For Mova (tokens): Before tokenization, a normalization pass could rejoin words split by line-break hyphens or broken extraction, so the tagger and lemmatizer see whole words.
  • I2. Remove small secondary marks (dots, diacritics) before finding the main units, so tiny pieces are not mistaken for units of their own. — For Mova (tokens): The tokenizer could set aside punctuation and similar attachments when finding word bodies, then reattach them deterministically, which keeps each step explainable.
  • I3. Pick an evaluation dataset whose ground truth labels the exact units being produced, so scoring can be fully automatic. — For Mova (learn): Scoring should prefer treebanks and test sets with gold annotation at the same granularity as Mova's output, instead of patching in hand-made alignment files.
  • I4. The number of expected pieces follows from a known rule (which letters do not join), giving a structural expectation to check output against. — For Mova (learn): Rule-derived expectations, such as the clause count implied by verbs and conjunctions, could act as cheap consistency checks that flag suspicious silver parses for repair.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.5121/sipij.2020.11602 · arXiv 2101.02797

@misc{abdallah2021off,
  title = {{Off-Line Arabic Handwritten Words Segmentation using Morphological Operators}},
  author = {AbdAllah, Nisreen and Viriri, Serestina},
  year = {2021},
  doi = {10.5121/sipij.2020.11602},
  eprint = {2101.02797},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.5121/sipij.2020.11602}
}

References

45 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. A. A. A. Ali and M. Suresha, "Survey on Segmentation and Recognition of Handwritten Arabic Script," 2019. find
  2. R. Plamondon and S. N. Srihari, "Online and off-line handwriting recognition: a comprehensive survey," IEEE Transactions on pattern analysis and machine intelligence, vol. 22, no. 1, pp. 63-84, 2000. find
  3. M. T. Parvez and S. A. Mahmoud, "Offline Arabic handwritten text recognition: a survey," ACM Computing Surveys (CSUR), vol. 45, no. 2, pp. 1-35, 2013. find
  4. M. Shatnawi, "Off-line handwritten Arabic character recognition: a survey," in Proceedings of the international conference on image processing, computer vision, and pattern recognition (IPCV), 2015: The Steering Committee of The World Congress in Computer Science, Computer …, p. 52. find
  5. A. M. Zeki, "The segmentation problem in arabic character recognition the state of the art," in 2005 International Conference on Information and Communication Technologies, 2005: IEEE, pp. 11-26. find
  6. I. Yousif and A. Shaout, "Off-Line handwriting Arabic text recognition: a survey," International Journal of Advanced Research in Computer Science and Software Engineering, vol. 4, no. 9, 2014. find
  7. S. Impedovo, L. Ottaviano, and S. Occhinegro, "Optical character recognition—a survey," International journal of pattern recognition and artificial intelligence, vol. 5, no. 01n02, pp. 1-24, 1991. find
  8. L. M. Lorigo and V. Govindaraju, "Offline Arabic handwriting recognition: a survey," IEEE transactions on pattern analysis and machine intelligence, vol. 28, no. 5, pp. 712-724, 2006. find
  9. C. C. Tappert, C. Y. Suen, and T. Wakahara, "The state of the art in online handwriting recognition," IEEE Transactions on pattern analysis and machine intelligence, vol. 12, no. 8, pp. 787-808, 1990. find
  10. M. S. Khorsheed, "Off-line Arabic character recognition–a review," Pattern analysis & applications, vol. 5, no. 1, pp. 31-45, 2002. 34 find
  11. L. Dinges, A. Al-Hamadi, M. Elzobi, Z. Al Aghbari, and H. Mustafa, "Offline automatic segmentation based recognition of handwritten arabic words," International Journal of Signal Processing, Image Processing and Pattern Recognition, vol. 4, no. 4, pp. 131-143, 2011. find
  12. F. B. Samoud, S. S. Maddouri, and H. Amiri, "Three Evaluation Criteria's towards a Comparison of Two Characters Segmentation Methods for Handwritten Arabic Script," in 2012 International Conference on Frontiers in Handwriting Recognition, 2012: IEEE, pp. 774-779. find
  13. H. Ghaleb, P. Nagabhushan, and U. Pal, "Segmentation of overlapped handwritten Arabic subwords," International Journal of Computer Applications, vol. 975, p. 8887, 2015. find
  14. A.-S. Atallah and K. Omar, "A comparative study between methods of Arabic baseline detection," in 2009 International Conference on Electrical Engineering and Informatics, 2009, vol. 1: IEEE, pp. 73-77. find
  15. Y. M. Alginahi, "A survey on Arabic character segmentation," International Journal on Document Analysis and Recognition (IJDAR), vol. 16, no. 2, pp. 105-126, 2013. find
  16. H. Almuallim and S. Yamaguchi, "A method of recognition of Arabic cursive handwriting," IEEE transactions on pattern analysis and machine intelligence, no. 5, pp. 715-722, 1987. find
  17. M. Cheriet, "Visual recognition of Arabic handwriting: challenges and new directions," in Summit on Arabic and Chinese Handwriting Recognition, 2006: Springer, pp. 1-21. find
  18. A. A. Aburas and M. E. Gumah, "Arabic handwriting recognition: Challenges and solutions," in 2008 International Symposium on Information Technology, 2008, vol. 2: IEEE, pp. 1-6. find
  19. P. Ahmed and Y. Al-Ohali, "Arabic character recognition: Progress and challenges," Journal of King Saud University-Computer and Information Sciences, vol. 12, pp. 85-116, 2000. find
  20. A. Lawgali, M. Angelova, and A. Bouridane, "A Frame Work For Arabic Handwritten Recognition Based on Segmentation," 2014. find
  21. D. Xiang, H. Yan, X. Chen, and Y. Cheng, "Offline Arabic handwriting recognition system based on HMM," in 2010 3rd International Conference on Computer Science and Information Technology, 2010, vol. 1: IEEE, pp. 526-529. find
  22. A. Farahmand, H. Sarrafzadeh, and J. Shanbehzadeh, "Document image noises and removal methods," 2013. find
  23. T. Abu-Ain, S. N. H. S. Abdullah, B. Bataineh, K. Omar, and A. Abu-Ein, "A novel baseline detection method of handwritten Arabic-script documents based on sub-words," in International Multi-Conference on Artificial Intelligence Technology, 2013: Springer, pp. 67-77. find
  24. F. Farooq, V. Govindaraju, and M. Perrone, "Pre-processing methods for handwritten Arabic documents," in Eighth International Conference on Document Analysis and Recognition (ICDAR'05), 2005: IEEE, pp. 267-271. find
  25. A. Baz and M. Baz, "A New Approach and Algorithm for Baseline Detection of Arabic Handwriting," 2015. find
  26. J. Song, R. L. Stevenson, and E. J. Delp, "The use of mathematical morphology in image enhancement," in Proceedings of the 32nd Midwest Symposium on Circuits and Systems, 1989: IEEE, pp. 67-70. find
  27. N. H. Barna, T. I. Erana, S. Ahmed, and H. Heickal, "Segmentation of heterogeneous documents into homogeneous components using morphological operations," in 2018 IEEE/ACIS 17th International Conference on Computer and Information Science (ICIS), 2018, pp. 513-518. find
  28. B. Goyal, A. Dogra, S. Agrawal, and B. Sohi, "Two-dimensional gray scale image denoising via morphological operations in NSST domain & bitonic filtering," Future Generation Computer Systems, vol. 82, pp. 158-175, 2018. find
  29. N. Jamil, T. M. T. Sembok, and Z. A. Bakar, "Noise removal and enhancement of binary images using morphological operations," in 2008 International Symposium on Information Technology, 2008, vol. 4: IEEE, pp. 1-6. find
  30. D. Motawa, A. Amin, and R. Sabourin, "Segmentation of Arabic cursive script," in Proceedings of the fourth international conference on document analysis and recognition, 1997, vol. 2: IEEE, pp. 625-628. find
  31. J. H. AlKhateeb, J. Jiang, J. Ren, and S. Ipson, "Component-based segmentation of words from handwritten Arabic text," International Journal of Computer Systems Science and Engineering, vol. 5, no. 1, 2009. find
  32. I. A. Humied, "Segmentation accuracy for offline Arabic handwritten recognition based on bounding box algorithm," International Journal of Computer Science and Network Security (IJCSNS), vol. 16, no. 9, p. 98, 2016. find
  33. M. Elzobi, A. Al-Hamadi, Z. Al Aghbari, and L. Dings, "IESK-ArDB: a database for handwritten Arabic and an optimized topological segmentation approach," International Journal on Document Analysis and Recognition (IJDAR), vol. 16, no. 3, pp. 295-308, 2013. find
  34. N. Arica and F. T. Yarman-Vural, "An overview of character recognition focused on off-line handwriting," IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), vol. 31, no. 2, pp. 216-233, 2001. find
  35. N. Otsu, "A threshold selection method from gray-level histograms," IEEE transactions on systems, man, and cybernetics, vol. 9, no. 1, pp. 62-66, 1979. find
  36. T. Zhang and C. Y. Suen, "A fast parallel algorithm for thinning digital patterns," Communications of the ACM, vol. 27, no. 3, pp. 236-239, 1984. find
  37. Y. Osman, "Segmentation algorithm for Arabic handwritten text based on contour analysis," in 2013 INTERNATIONAL CONFERENCE ON COMPUTING, ELECTRICAL AND ELECTRONIC ENGINEERING (ICCEEE), 2013: IEEE, pp. 447-452. find
  38. R. G. Casey and E. Lecolinet, "A survey of methods and strategies in character segmentation," IEEE transactions on pattern analysis and machine intelligence, vol. 18, no. 7, pp. 690-706, 1996. find
  39. A. Rosenfeld and J. Pfaltz, "Sequential Operations in Digital Picture Processing; JACM 13 (1966) 471–494," Google Scholar Google Scholar Digital Library Digital Library, 1966. find
  40. H. Samet and M. Tamminen, "An improved approach to connected component labeling of images," in International Conference on Computer Vision And Pattern Recognition, 1986, vol. 318, p. 312. find
  41. L. Di Stefano and A. Bulgarelli, "A simple and efficient connected components labeling algorithm," in Proceedings 10th International Conference on Image Analysis and Processing, 1999: IEEE, pp. 322-327. find
  42. L. Berriche and A. Al-Mutairy, "Seam carving-based Arabic handwritten sub-word segmentation," Cogent Engineering, vol. 7, no. 1, p. 1769315, 2020. find
  43. N. Essa, E. El‐Daydamony, and A. A. Mohamed, "Enhanced technique for Arabic handwriting recognition using deep belief network and a morphological algorithm for solving ligature segmentation," ETRI Journal, vol. 40, no. 6, pp. 774-787, 2018. find
  44. A. Q. M. S. Zerdoumi Saber, A. Kamsin, and S. Hakak, "Efficient Approach to Segment Ligatures and Open Characters in Offline Arabic text," 2017. find
  45. L. S. Alhomed and K. M. Jambi, "A Survey on the Existing Arabic Optical Character Recognition and Future Trends," International Journal of Advanced Research in Computer and Communication Engineering (IJARCCE), vol. 7, no. 3, pp. 78-88, 2018. A UTHORS NISREEN ABDALLAH received a BSc degree in computer science from Omdurman Islamic University, and an M.Sc. degree from the University of Khartoum, Sudan. She is currently pursuing a Ph.D. degree in computer science at Sudan University of Science and Technology, Sudan. She also works as a lecture in the Department of Information Technology at the University of Kordofan. Her research interests include machine learning, computer vision, image processing, and pattern recognition. She attended and presented in some national and international conferences about handwriting recognition. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/abdallah-2021-off-line-arabic-handwritten-words/