Skip to main content
SpringerLink
Log in

An image database of handwritten Bangla words with automatic benchmarking facilities for character segmentation algorithms

  • Original Article
  • Published:
Neural Computing and Applications Aims and scope Submit manuscript

Abstract

Recognition of unconstrained handwritten word images is an interesting research problem which gets more challenging when lexicon-free words are considered. Prerequisite for developing a lexicon-free handwritten word recognition technique is the segmentation of a word image into its constituent character set. Therefore, a competent character segmentation technique is required to design a comprehensive word recognition module. However, the literature study reveals that there is no standard word image database with ground truth information. As a result, most character segmentation algorithms found in the literature rely on self-made databases with manual evaluation. To fill the research need, in the present scope of the work, a comprehensive database consisting of handwritten Bangla word images is prepared primarily for evaluating any character segmentation algorithms. Additionally, the present work also provides two types of ground truth images related to segmented character shapes of the word images. Besides, an evaluation tool is developed for assessing the performance of any character segmentation algorithm on the developed benchmark database. The benchmark result, as found here, is 0.9212 (F-score) which outperforms some state-of-the-art methods.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9

Similar content being viewed by others

References

  1. He S, Schomaker L (2019) DeepOtsu: document enhancement and binarization using iterative deep learning. Pattern Recognit 91:379–390

    Article  Google Scholar 

  2. Dixit UD, Shirdhonkar MS (2019) Preprocessing framework for document image analysis. Int J Adv Netw Appl 10(4):3911–3918

    Google Scholar 

  3. Bera SK, Chakrabarti A, Lahiri S, Smith EHB, Sarkar R (2019) Normalization of unconstrained handwritten words in terms of Slope and Slant Correction. Pattern Recognit Lett 128:488–495

    Article  Google Scholar 

  4. Kundu S, Paul S, Bera SK, Abraham A, Sarkar R (2020) Text-line extraction from handwritten document images using GAN. Expert Syst Appl 140:112916

    Article  Google Scholar 

  5. Sarkar R, Halder S, Malakar S, Das N, Basu S, Nasipuri M (2012) Text line extraction from handwritten document pages based on line contour estimation. In: 2012 3rd international conference on computing, communication and networking technologies, ICCCNT 2012

  6. Sarkar R, Moulik S, Das N, Basu S, Nasipuri M, Basu DK (2011) Word extraction from unconstrained handwritten Bangla document images using Spiral Run Length Smearing Algorithm. In: Proceedings of the 5th Indian international conference on artificial intelligence, IICAI 2011, pp 32–46

  7. Bhowmik S, Malakar S, Sarkar R, Basu S, Kundu M, Nasipuri M (2019) Off-line Bangla handwritten word recognition: a holistic approach. Neural Comput Appl 31(10):5783–5798

    Article  Google Scholar 

  8. Dhaka VP, Sharma MK (2015) An efficient segmentation technique for Devanagari offline handwritten scripts using the Feedforward Neural Network. Neural Comput Appl 26(8):1881–1893

    Article  Google Scholar 

  9. Mellouli D, Hamdani TM, Sanchez-Medina JJ, Ben Ayed M, Alimi AM (2019) Morphological convolutional neural network architecture for digit recognition. IEEE Trans Neural Netw Learn Syst

  10. Mozafari M, Ganjtabesh M, Nowzari-Dalini A, Thorpe SJ, Masquelier T (2019) Bio-inspired digit recognition using reward-modulated spike-timing-dependent plasticity in deep convolutional networks. Pattern Recognit 94:87–95

    Article  Google Scholar 

  11. Guha R, Das N, Kundu M, Nasipuri M, Santosh KC (2019) DevNet: an efficient CNN architecture for handwritten Devanagari character recognition. Int J Pattern Recognit Artif Intell

  12. Malakar S, Ghosh M, Bhowmik S, Sarkar R, Nasipuri M (xxxx) A GA based hierarchical feature selection approach for handwritten word recognition. Neural Comput Appl

  13. Malakar S, Sharma P, Singh PK, Das M, Sarkar R, Nasipuri M (2017) A holistic approach for handwritten hindi word recognition. Int J Comput Vis Image Process 7(1):59–78

    Article  Google Scholar 

  14. Sahoo S et al (2018) Handwritten Bangla word recognition using negative refraction based shape transformation. J Intell Fuzzy Syst 35(2):569

    Google Scholar 

  15. Wu X, Chen Q, You J, Xiao Y (2019) Unconstrained offline handwritten word recognition by position embedding integrated ResNets model. IEEE Signal Process Lett 26(4):597–601

    Article  Google Scholar 

  16. Gui L, Liang X, Chang X, Hauptmann AG (2018) Adaptive context-aware reinforced agent for handwritten text recognition. In: BMVC, p 207

  17. Tulyakov S, Govindaraju V (2001) Probabilistic model for segmentation based word recognition with lexicon. In: Proceedings of sixth international conference on document analysis and recognition, pp 164–167

  18. El-Yacoubi A, Gilloux M, Sabourin R, Suen CY (1999) An HMM-based approach for off-line unconstrained handwritten word modeling and recognition. IEEE Trans Pattern Anal Mach Intell 21(8):752–760

    Article  Google Scholar 

  19. Vinciarelli A (2002) A survey on off-line cursive word recognition. Pattern Recognit 35(7):1433–1446

    Article  Google Scholar 

  20. Zhang C, Li W (2013) Recognizing handwritten Chinese day and month words by combining a holistic method and a segmentation-based method. Neural Comput Appl 23(6):1661–1668

    Article  Google Scholar 

  21. Kim KK, Kim JH, Chung YK, Suen CY (2001) Legal amount recognition based on the segmentation hypotheses for bank check processing. In: Proceedings of sixth international conference on document analysis and recognition, pp 964–967

  22. Roy PP, Bhunia AK, Das A, Dey P, Pal U (2016) HMM-based Indic handwritten word recognition using zone segmentation. Pattern Recognit 60:1057–1075

    Article  Google Scholar 

  23. Malakar S, Ghosh P, Sarkar R, Das N, Basu S, Nasipuri M (2011) An improved offline handwritten character segmentation algorithm for Bangla script. In: Proceedings of the 5th Indian international conference on artificial intelligence, IICAI 2011

  24. Bhowmik S, Polley S, Roushan MG, Malakar S, Sarkar R, Nasipuri M (2015) A holistic word recognition technique for handwritten Bangla words. Int J Appl Pattern Recognit 2(2):142–159

    Article  Google Scholar 

  25. Basu S, Das N, Sarkar R, Kundu M, Nasipuri M, Basu DK (2009) A hierarchical approach to recognition of handwritten Bangla characters. Pattern Recognit 42(7):1467–1484

    Article  Google Scholar 

  26. Sarkar R, Malakar S, Das N, Basu S, Kundu M, Nasipuri M (2011) Word extraction and character segmentation from text lines of unconstrained handwritten Bangla document images. J Intell Syst 20(3):227–260

    Google Scholar 

  27. Sarkar R, Malakar S, Das N, Basu S, Nasipuri M (2010) A script independent technique for extraction of characters from handwritten word images. Int J Comput Appl 1(23):83–88

    Google Scholar 

  28. Sarkar R, Sen B, Das N, Basu S (2008) Handwritten devanagari script segmentation: a non-linear fuzzy approach, arXiv1501.05472

  29. Hull JJ (1994) A database for handwritten text recognition research. IEEE Trans Pattern Anal Mach Intell 16(5):550–554

    Article  Google Scholar 

  30. Wilkinson RA et al. (1992) The first census optical character recognition system conference, vol. 184. US Department of Commerce, National Institute of Standards and Technology

  31. ICDAR 2011 Competition. [Online]. http://www.icdar2011.org/EN/column/column26.shtml

  32. CMATERdb2.1.1. [Online]. https://drive.google.com/open?id=0B8usn8guzQEMVDdwYW4zRTFnRm8

  33. Sarkar R, Das N, Basu S, Kundu M, Nasipuri M, Basu DK (2012) CMATERdb1: a database of unconstrained handwritten Bangla and Bangla-English mixed script document image. Int J Doc Anal Recognit 15(1):71–83

    Article  Google Scholar 

  34. Otsu N (1979) A threshold selection method from gray-level histograms. IEEE Trans Syst Man Cybern 9(1):62–66

    Article  Google Scholar 

  35. Berg R (xxxx) Sensitivity and specificity. https://en.wikipedia.org/wiki/Sensitivity_and_specificity. Accessed on 17 Dec 2018

  36. Basu S, Sarkar R, Das N, Kundu M, Nasipuri M, Basu DK (2007) A fuzzy technique for segmentation of handwritten Bangla word images. Proc Int Conf Comput Theory Appl ICCTA 2007:427–432

    Google Scholar 

  37. Singh PK, Mahanta S, Malakar S, Sarkar R, Nasipuri M (2014) Development of a page segmentation technique for Bangla documents printed in italic style. In: 2014 2nd international conference on business and information management, ICBIM 2014

  38. Roy A, Bhowmik TK, Parui SK, Roy U (2005) A novel approach to skew detection and character segmentation for handwritten Bangla words. In: Proceedings of the digital imaging computing: techniques and applications, DICTA 2005, vol 2005, pp 203–210

  39. Yamaguchi T et al. (2002) A segmentation system for touching handwritten Japanese characters. In: Proceedings eighth international workshop on frontiers in handwriting recognition, pp 407–412

  40. Roy PP, Pal U, Llados J, Delalandre M (2009) Multi-oriented and multi-sized touching character segmentation using dynamic programming. In: 2009 10th international conference on document analysis and recognition, pp 11–15

  41. Sas J, Markowska-Kaczmar U (2007) Semi-supervised handwritten word segmentation using character samples similarity maximization and evolutionary algorithm. In: Proceedings—6th international conference on computer information systems and industrial management applications, CISIM 2007, pp 316–321

  42. Sharma DV, Lehal GS (2006) An iterative algorithm for segmentation of isolated handwritten words in Gurmukhi script. Proc Int Conf Pattern Recognit 2:1022–1025

    Google Scholar 

Download references

Acknowledgements

We would like to thank CMATER research laboratory of the Computer Science and Engineering Department, Jadavpur University, India, for providing us the infrastructural support. This work is partially supported by the PURSE-II and UPE-II, Jadavpur University projects. Ram Sarkar is thankful to DST, Govt. of India, for the grant (EMR/2016/007213) to carry out this research.

Author information

Authors and Affiliations

  1. Department of Computer Science, Asutosh College, Kolkata, India

    Samir Malakar

  2. Department of Computer Science and Engineering, Jadavpur University, Kolkata, India

    Ram Sarkar, Subhadip Basu, Mahantapas Kundu & Mita Nasipuri

Authors
  1. Samir Malakar
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Ram Sarkar
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Subhadip Basu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Mahantapas Kundu
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Mita Nasipuri
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Samir Malakar.

Ethics declarations

Conflict of interest

We declare that we have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

1.1 Estimation of actual dimension of word image

Estimation of minimum rectangular bounding box (Fig. 10) covering the entire word is carried out to measure word dimension. Let \( StartR, StartC, EndR, EndC\) are the starting row, starting column, ending row, and ending column of a word image, respectively. Actual height (\(Ht\)) and width (\(Wd\)) of the image are calculated by \(Ht = ER - SR + 1\) and \(Wd = EC - SC + 1\).

Fig. 10
figure 10

Depiction of word boundary after (bold outer boundary) and before (dotted outer boundary) cropping out the unwanted region

Full size image

See Table 11.

Table 11 Description of notations used to define word images
Full size table

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Malakar, S., Sarkar, R., Basu, S. et al. An image database of handwritten Bangla words with automatic benchmarking facilities for character segmentation algorithms. Neural Comput & Applic 33, 449–468 (2021). https://doi.org/10.1007/s00521-020-04981-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00521-020-04981-w

Keywords

Search

Navigation