A Systematic Framework for Sanskrit Character Recognition Using Deep Learning
Abstract
Sanskrit is widely acknowledged to be among the world’s oldest surviving classical languages, and yet its usage has continued to decline unabated in the present milieu. Such insidious erosion of popularity is directly attributable to the absence of native speakers of the language and the perceived inaccessibility of Sanskrit to contemporary audiences. Notwithstanding, the language remains historically and culturally inseparable from the subcontinent, with numerous religious manuscripts, epigraphical inscriptions, edicts and scientific literature written in the Sanskrit script. Attempts made to resuscitate the language have been largely unsuccessful as these attempts have relied extensively on laborious human transcription and translation. Such manual endeavors can be superseded by the use of efficient computational techniques to facilitate the efficient transcription of voluminous manuscripts written in the Sanskrit script.
The emergence of deep learning frameworks has enabled researchers to overcome the draw backs of conventional machine learning algorithms in developing efficient and extensible character recognition systems. Notwithstanding, the advancement of character recognition frameworks varies across different Indic scripts.
In this context, this paper introduces an extensible framework for the transcription of hand written Sanskrit manuscripts. In the absence of a benchmark dataset of handwritten Sanskrit characters, the authors introduce a comprehensive dataset to facilitate further downstream segmentation. The dataset, on augmentation, comprises over a hundred thousand samples and has been collected from over a hundred individuals. The paper explores an integrated approach to segmentation and accordingly delineates a systematic methodology for effectively segmenting Sanskrit words, incorporating techniques such as thresholding, zone-based classification, median bisection and projection profiles. The proposed technique accommodates a diverse array of characters and modifiers present in the Sanskrit script. Subsequently, a concurrent deep learning architecture parallelizes transcription using Neural Networks (CNN and Residual Networks). The deep learning models show accuracies exceeding 90%. This paper attempts to benchmark the significance of
systematic approaches to machine transcription of low-resource languages.
Keywords
Sanskrit, Devanagari, low-resource languages, transcription, optical character recognition, segmentation, deep learning, neural networksReferences
[1] V. A. Raj, R. L. Jyothi and A. Anilkumar, ”Grantha script recognition from ancient palm leaves using histogram of orientation shape context,” 2017 International Conference on Computing Methodologies and Communication (ICCMC), Erode, India, 2017, pp. 790-794, doi: 10.1109/IC
CMC.2017.8282574.
[2] A. Tripathi, S. S. Paul and V. K. Pandey, ”Standardisation of stroke order for online isolated Devanagari character recognition for iPhone,” 2012 IEEE International Conference on Technology Enhanced Education (ICTEE), Amritapuri, India, 2012, pp. 1-5, doi: 10.1109/ICTEE.2012.6208657.
[3] N. Sethi, A. Dev and P. Bansal, ”A Bilingual Machine Transliteration System for Sanskrit-English Using Rule-Based Approach,” 2022 4th International Conference on Artificial Intelligence and Speech Technology (AIST), Delhi, India, 2022, pp. 1-5, doi: 10.1109/AIST55798.2022.10064993.
[4] A. Das, A. S. Azad Rabby, I. Kowsar and F. Rahman, ”A Deep Learning-based Unified Solution for Character Recognition,” 2022 26th International Conference on Pattern Recognition (ICPR), Montreal, QC, Canada, 2022, pp. 1671-1677, doi: 10.1109/ICPR56361.2022.9956348.
[5] J. Mehta and N. Garg, ”Offline Handwritten Sanskrit Simple and Compound Character Recognition Using Neural Network,” in Proceedings of International Conference on ICT for Sustainable Development, vol. 408, Springer, Singapore, 2016, pp. 62.
[6] N. Palrecha, A. Rai, A. Kumar, S. Srivastava and V. Tyagi, ”Character segmentation for multilingual Indic and Roman scripts,” 2011 IEEE 7th International Colloquium on Signal Processing and its Applications, Penang, Malaysia, 2011, pp. 45-49, doi: 10.1109/CSPA.2011.5759840.
[7] M. Sonkusare, R. Gupta, and A. Moghe, ”A Review on Character Segmentation Approach for Devanagari Script,” in Intelligent Systems, Algorithms for Intelligent Systems, Springer, Singapore, 2021, pp. 19.
[8] S. Kumar, ”A study for handwritten Devanagari word recognition,” 2016 International Conference on Communication and Signal Processing (ICCSP), Melmaruvathur, India, 2016, pp. 1009-1014, doi: 10.1109/ICCSP.2016.7754301.
[9] A. Dwivedi, R. Saluja and R. K. Sarvadevabhatla, ”An OCR for Classical Indic Documents Containing Arbitrarily Long Words,” 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, WA, USA, 2020, pp. 2386-2393, doi:
10.1109/CVPRW50498.2020.00288.
[10] Gadade, Y. A. (2019). Improving Accuracy of the Tesseract. limitlessdatascience.wordpress.com. https://limitlessdatascience.wordpress.com/2019/05/01/improving-accuracy-of-the-tesseract/ (Accessed Feb. 23 2024)
[11] R. Shah, M. K. Gupta and A. Kumar, ”Ancient Sanskrit Line-level OCR using OpenNMT Architecture,” 2021 Sixth International Conference on Image Information Processing (ICIIP), Shimla, India, 2021, pp. 347-352, doi: 10.1109/ICIIP53038.2021.9702666.
[12] M. Avadesh and N. Goyal, ”Optical Character Recognition for Sanskrit Using Convolution Neural Networks,” 2018 13th IAPR International Workshop on Document Analysis Systems (DAS), Vienna, Austria, 2018, pp. 447-452, doi: 10.1109/DAS.2018.50.
[13] C. Halder and K. Roy, ”Word & Character Segmentation for Bangla Handwriting Analysis & Recognition,” 2011 Third National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics, Hubli, India, 2011, pp. 243-246, doi: 10.1109/NCVPRIPG.2011.68.
[14] Y. Gurav, P. Bhagat, R. Jadhav and S. Sinha, ”Devanagari Handwritten Character Recognition using Convolutional Neural Networks,” 2020 International Conference on Electrical, Communication, and Computer Engineering (ICECCE), Istanbul, Turkey, 2020, pp. 1-6, doi:
10.1109/ICECCE49384.2020.9179193.
[15] K. C. Santosh, C. Nattee and B. Lamiroy, ”Spatial Similarity Based Stroke Number and Order Free Clustering,” 2010 12th International Conference on Frontiers in Handwriting Recognition, Kolkata, India, 2010, pp. 652-657, doi: 10.1109/ICFHR.2010.107.
[16] S. Singh and M. Sachan, ”Opportunities and Challenges of Handwritten Sanskrit Character Recognition System,” in International Journal of Computer Applications, vol. 3, pp. 18-22, 2017.
[17] A. Torralba and A. A. Efros, ”Unbiased look at dataset bias,” CVPR 2011, Colorado Springs, CO, USA, 2011, pp. 1521-1528, doi: 10.1109/CVPR.2011.5995347.
[18] Z. Zhao and Q. Ma, ”A novel method for image clustering,” 2014 10th International Conference on Natural Computation (ICNC), Xiamen, China, 2014, pp. 648-652, doi: 10.1109/ICNC.2014.6975912.
[19] A. Alkandari and S. J. Aljaber, ”Principle Component Analysis algorithm (PCA) for image recognition,” 2015 Second International Conference on Computing Technology and Information Management (ICCTIM), Johor, Malaysia, 2015, pp. 76-80, doi: 10.1109/ICCTIM.2015.7224596.
[20] K. P. Sinaga and M.-S. Yang, ”Unsupervised K-Means Clustering Algorithm,” in IEEE Access, vol. 8, pp. 80716-80727, 2020, doi: 10.1109/ACCESS.2020.2988796.
[21] A. K. Bathla, S. K. Gupta and M. K. Jindal, ”Challenges in recognition of Devanagari Scripts due to segmentation of handwritten text,” 2016 3rd International Conference on Computing for Sustainable Global Development (INDIACom), New Delhi, India, 2016, pp. 2711-2715.
[22] N. K. Garg, L. Kaur and M. K. Jindal, ”A New Method for Line Segmentation of Handwritten Hindi Text,” 2010 Seventh International Conference on Information Technology: New Generations, Las Vegas, NV, USA, 2010, pp. 392-397, doi: 10.1109/ITNG.2010.89.
[23] M. I. Bhat, B. Sharada, S. M. Obaidullah and M. Imran, ”Towards Accurate Identification and Removal of Shirorekha from Off-line Handwritten Devanagari word Documents,” 2020 17th International Conference on Frontiers in Handwriting Recognition (ICFHR), Dortmund, Germany,
2020, pp. 234-239, doi: 10.1109/ICFHR2020.2020.00051.
[24] A. B. Shinde and Y. H. Dandawate, ”Shirorekha extraction in Character Segmentation for printed devanagri text in Document Image Processing,” 2014 Annual IEEE India Conference (INDICON), Pune, India, 2014, pp. 1-7, doi: 10.1109/INDICON.2014.7030535.
[25] R. Ghosh and P. P. Roy, ”Comparison of Zone-Features for Online Bengali and Devanagari Word Recognition Using HMM,” 2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR), Shenzhen, China, 2016, pp. 435-440, doi: 10.1109/ICFHR.2016.0087.
[26] N. Aneja and S. Aneja, ”Transfer Learning using CNN for Handwritten Devanagari Character Recognition,” 2019 1st International Conference on Advances in Information Technology (ICAIT), Chikmagalur, India, 2019, pp. 293-296, doi: 10.1109/ICAIT47043.2019.8987286.
[27] S. Patnaik, S. Kumari and S. Das Mahapatra, ”Comparison of deep CNN and ResNet for Handwritten Devanagari Character Recognition,” 2020 IEEE 1st International Conference for Convergence in Engineering (ICCE), Kolkata, India, 2020, pp. 235-238, doi:
10.1109/ICCE50343.2020.9290637.
[28] S. S. Magare and R. R. Deshmukh, ”Offline Handwritten Sanskrit Character Recognition Using Hough Transform and Euclidean Distance,” International Journal of Innovation and Scientific Research, vol. 10, no. 2, pp. 295-302, Oct. 2014
Published
How to Cite
Downloads
Copyright (c) 2025 Vrinda Kore, Dhruva G, Sahana Rao, Vijitha M, Preethi
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
