Digitally Preserving Devnagri Script by using deep learning
DOI:
https://doi.org/10.63503/j.ijcma.2025.189Keywords:
Devanagari script, handwritten character recognition, deep convolutional neural network (DCNN), image preprocessing, digitization, optical character recognition (OCR)Abstract
Devanagari is one of the most widely used and ancient notation systems in South Asia, primarily used for languages like Hindi, Sanskrit, Marathi, and Nepali. Devanagari is used to write ancient languages of the Hindu Holy Writ, philosophy, and classical literature. It preserves the centuries of India's cultural, religious, and it’s intellectual heritage. Some handbooks like the Bhagavad Gita, Vedas, and Upanishads are also available in Devanagari. Therefore, it's our responsibility that we bring Devanagari in digital form. This paper presents a new system to digitize the handwritten Images based on image processing methods along with Deep Convolutional Neural Network (DCNN). Handwritten character recognition in Devanagari presents significant challenges due to the script’s complex structure, numerous character classes. To address these limitations, the paper introduces an approach that uses combination of image preprocessing ways and a Deep Convolutional Neural Network (DCNN) for point birth and type. The proposed system enhances input images through preprocessing way analogous as grayscale conversion, noise reduction, binarization, and segmentation to use a DCNN trained specifically for Devanagari characters to prize features and classify them directly from analogous images. To meliorate contextual understanding, a secondary model reconstructs meaningful words and rulings from the recognized characters, icing verbal consonance. This double- model frame — combining visual recognition with language modelling — enables robust digitization of handwritten Devanagari text. The system aims to save nonfictional calligraphies and grease digital archiving of Devanagari handwritten images.
References
[1] S. Arora, L. Malik, S. Goyal, D. Bhattacharjee, M. Nasipuri, and O. Krejcar, “Devanagari character recognition: A comprehensive literature review,” IEEE Access, vol. 13, pp. 1249–1284, 2025, doi: 10.1109/ACCESS.2024.3520248.
[2] K. Thakur and S. Phadikar, “Writer identification from handwritten Devanagari script,” in Advances in Intelligent Systems and Computing, Springer, 2015, pp. 497–505, doi: 10.1007/978-81-322-2247-7_51.
[3] H. Sharma and A. Mohan Gupta, “Comprehensive survey on Devanagari OCR,” SSRN preprint, 2022. [Online]. Available: https://ssrn.com/abstract=4033489
[4] M. Noonan, “Recent adaptations of the Devanagari script for the Tibeto-Burman languages of Nepal,” 2010.
[5] T. Kundaikar and J. D. Pawar, “Multi-font Devanagari text recognition using LSTM neural networks,” in Advances in Intelligent Systems and Computing, Springer, 2020, pp. 495–506, doi: 10.1007/978-981-15-0029-9_39.
[6] R. Ahmad, “Urdu in Devanagari: Shifting orthographic practices and Muslim identity in Delhi,” Language in Society, vol. 40, no. 3, pp. 259–284, Jun. 2011, doi: 10.1017/S0047404511000182.
[7] O. V. Ramana Murthy, S. Roy, V. Narang, and M. Hanmandlu, “Devanagari character recognition in the wild,” International Journal of Computer Applications, vol. 38, no. 4, pp. 38–45, Jan. 2012, doi: 10.5120/4599-6800.
[8] S. Puri and S. P. Singh, “Efficient Devanagari character classification in printed and handwritten documents using SVM,” Procedia Computer Science, vol. 152, pp. 111–121, 2019, doi: 10.1016/j.procs.2019.05.033.
[9] S. Chanda and U. Pal, “English, Devanagari, and Urdu text identification,” Indian Statistical Institute, Kolkata, India.
[10] S. Kumar, “An analysis of irregularities in Devanagari script writing: A machine recognition perspective,” 2010.
[11] V. Kumar, R. Pankaj, and K. Sengar, “Segmentation of printed text in Devanagari and Gurmukhi scripts,” 2010.
[12] S. Bag and G. Harit, “A survey on optical character recognition for Bangla and Devanagari scripts”, 2013.
[13] S. Khare and J. Singh, “Handwritten Devanagari character recognition system: A review” 2015.
[14] R. Mhaiskar, “Devanagari: A historical overview and a study of writing style in Aga Khan Palace photo gallery,” Bulletin of the Deccan College Research Institute, vol. 74, pp. 197–208, 2014, doi: 10.2307/26264698.
[15] S. Musa, “Nayanagari: A simple script for Devanagari,” unpublished manuscript, 2006.
[16] R. Ahmad, “Scripting a new identity: The battle for Devanagari in nineteenth-century India,” Journal of Pragmatics, vol. 40, no. 7, pp. 1163–1183, Jul. 2008, doi: 10.1016/j.pragma.2007.06.005.
[17] R. Snell, “Transliterating Devanagari,” unpublished instructional material.
[18] N. Sankaran, A. Neelappa, and C. V. Jawahar, “Devanagari text recognition: A transcription-based formulation,” unpublished conference manuscript, 2013.
[19] T. Karayil, A. Ul-Hasan, and T. M. Breuel, “A segmentation-free approach for printed Devanagari script recognition,” in Proc. Int. Conf. Document Analysis and Recognition (ICDAR), Nov. 2015, pp. 946–950, doi: 10.1109/ICDAR.2015.7333901.
[20] M. S. Qureshi, M. S. Shaik, M. R. Murthy, and R. M. Saqib, “Devanagari character recognition: A short review,” unpublished review paper, 2012.
[21] N. C. Singh and C. Rao, “Reading in Devanagari: Insights from functional neuroimaging,” Indian Journal of Radiology and Imaging, vol. 24, no. 1, pp. 44–50, Feb. 2014, doi: 10.4103/0971-3026.130691.
[22] T. Das, U. Kumar, R. S. Bapi, P. Padakannaya, and N. C. Singh, “Neural representation of an alpha syllabary: The story of Devanagari,” unpublished research article, 2009.
[23] S. D. Pande et al., “Digitization of handwritten Devanagari text using CNN transfer learning,” NeuroImage: Reports, vol. 2, 2022, doi: 10.1016/j.neuri.2021.100016.
[24] S. D. Connell, R. M. K. Sinha, and A. K. Jain, “Recognition of unconstrained online Devanagari characters,” in Proc. 9th Int. Conf. Software, Knowledge, Information Management and Applications (SKIMA), IEEE, 2015.
[25] R. Jayadevan, S. R. Kolhe, P. M. Patil, and U. Pal, “Offline recognition of Devanagari script: A survey,” IEEE Trans. Systems, Man, and Cybernetics, Part C, vol. 41, no. 6, pp. 782–796, 2011, doi: 10.1109/TSMCC.2010.2095841.
[26] J. Vaid and A. Gupta, “Exploring word recognition in a semi-alphabetic script: The case of Devanagari,” Brain and Language, vol. 81, no. 1–3, pp. 679–690, 2002, doi: 10.1006/brln.2001.2556.
[27] S. Yalamanchili and K. Rao, “A framework for Devanagari script-based CAPTCHA”.
[28] S. Narang, M. K. Jindal, and M. Kumar, “Devanagari ancient documents recognition using statistical feature extraction techniques,” Sādhanā, vol. 44, no. 6, 2019, doi: 10.1007/s12046-019-1126-9.
[29] N. Sankaran and C. V. Jawahar, “Recognition of printed Devanagari text using BLSTM neural networks,”.
