The increasing prevalence of automated bots on the internet has highlighted the need for robust CAPTCHA systems to protect online platforms from malicious activities. This paper focuses on the recognition of numerals in historical Gurmukhi manuscripts and presents a novel application of secure CAPTCHA generation with blockchain integration. The work begins by developing a dataset of numerals extracted from historical Gurmukhi manuscripts and performing preprocessing techniques. Various feature extraction methods are explored and machine learning classifiers are employed for evaluation. The proposed recognition model achieves a notable recall rate of 93.47% using SVM with RBF kernel on the combined features, demonstrating its efficacy in recognizing unique characters from historical Gurmukhi manuscripts. Building on the successful numeral recognition model, the paper introduces the application of secure CAPTCHA generation using numerals of historical Gurmukhi manuscript. In this approach, references or hashes of the trained model parameters and dataset are securely stored on a blockchain, ensuring transparent access to the data and model. During the CAPTCHA generation phase, random images are selected from the blockchain dataset and recognized using the trained model. The recognition results are securely stored on the blockchain. These images are then distorted and presented to users in a challenging format. User responses are recorded and crossreferenced with the recognition results on the blockchain, enabling verification of human-generated responses and distinguishing them from automated bot activity. The integration of numeral recognition from historical Gurmukhi manuscripts with secure CAPTCHA generation and blockchain technology offers a promising solution to enhance online security.

Aggarwal, A., and Singh, S. (2012), “Offline handwritten Gurumukhi numeral recognition using SVM and different feature set”, International Journal of Advanced Research in Computer Science, 3(3):1–5.

Aggarwal, A., Singh, K.., and Singh, K.. (2015). “Use of gradient technique for extracting features from handwritten Gurmukhi characters and numerals.” International conference on Information and Communication Technologies (ICICT 2014). Dec 3-5. Kochi, India.

Arora, S., Bhattacharjee, D., Nasipuri, M., Basu, K., D. and Kundu, M. (2008). “Combining multiple feature extraction techniques for handwritten Devnagari character recognition”. 3rd International Conference on Industrial and Information Systems (ICIIS). Dec 8-10. Kharagpur, India.

Chacko, B. P. and Anto, B. P. (2010). “Pre and Post processing approaches in edge detection for character recognition”.12th International Conference on the Frontiers of Handwriting Recognition (ICFHR). Nov 16-18. Kolkata, India.

Kaur, K., Chaudhuri, B.B., and Lehal, G.S. (2022), 'A Benchmark Gurmukhi Handwritten Character Dataset: Acquisition, Compilation, and Recognition', in Frontiers in Handwriting Recognition. ICFHR 2022, eds. U. Porwal, A. Fornés, and F. Shafait, Springer. https://doi.org/10.1007/978-3-031-21648-0_31

Kumar, M., Jindal, M.K., and Kumar, M. (2021),” A Systematic Survey on CAPTCHA Recognition: Types, Creation and Breaking Techniques”, Archives of Computational Methods in Engineering, 29: 1107 - 1136.

Kumar, M., Jindal, M.K. and Kumar, M.(2022), “Design of innovative CAPTCHA for hindi language”, Neural Computing and Applications,34: 4957–4992. https://doi.org/10.1007/s00521-021-06686-0

Kumar, M., Jindal, M. K., Sharma, R. K., and Jindal, S. R. (2018), “Offline handwritten numeral recognition using combination of different feature extraction techniques”, National Academy Science Letters, 41: 29-33.

Kumar, M., Jindal, M. K., Sharma, R. K., and Jindal, S. R. (2020), “Performance evaluation of classifiers for the recognition of offline handwritten Gurmukhi characters and numerals: a study”, Artificial Intelligence Review, 53: 2075-2097. https://doi.org/10.1007/s10462-019-09727-2

Mahto, M. K., Bhatia, K., and Sharma, R. K. (2021), “Deep learning based models for offline Gurmukhi handwritten character and numeral recognition”, ELCVIA Electronic Letters on Computer Vision and Image Analysis, 20(2): 69-82.

Pradeep, J., Srinivasan, E., and Himavathi, S. (2011). “Diagonal based feature extraction for handwritten character recognition system using neural network”. 3rd International Conference on Electronics Computer Technology. April 8-10. Kanyakumari, India.

Sarangi, P.K., Sahoo, A.K., Kaur, G., Nayak, S.R., and Bhoi, A.K. (2022), 'Gurmukhi Numerals Recognition Using ANN', in Cognitive Informatics and Soft Computing, eds. P.K. Mallick, A.K. Bhoi, P. Barsocchi, and V.H.C. de Albuquerque, Springer. https://doi.org/10.1007/978-981-16-8763-1_30

Sarangi, P.K., Sahoo, A.K., Nayak, S.R., Agarwal, A., and Sethy, A. (2022), 'Recognition of Isolated Handwritten Gurumukhi Numerals Using Hopfield Neural Network', in Computational Intelligence in Pattern Recognition, eds. A.K. Das, J. Nayak, B. Naik, S. Dutta, and D. Pelusi, Springer. https://doi.org/10.1007/978-981-16-2543-5_51

Shahnaz, A., Qamar, U., and Khalid, A. (2019), “Using blockchain for electronic health records”, IEEE access, 7: 147782-147795.

Siddharth, K. S., Dhir, R., Rani, R., Jangid, M., and Singh, K. (2012).”Comparative Recognition of Handwritten Gurmukhi Numerals Using Different Feature Sets and Classifiers”. International Conference on Recent Advances and Future Trends in Information Technology (iRAFIT2012).March 21-23. Punjab, India.

Siddharth, K. S., Jangid, M., Dhir, R., and Rani, R. (2011), “Handwritten Gurmukhi character recognition using statistical and background directional distribution”, International Journal of Computer Science Engineering, 3(6): 2332-2345.

Singh, P., and Budhiraja, S. (2012), “Offline handwritten Gurmukhi numeral recognition using wavelet transforms”, International Journal of Modern Education and Computer Science, 4(8): 34.

Singh, S., and Dhir, R. (2012), “Recognition of Handwritten Gurmukhi Numeral using Gabor Filters”, International Journal of Computer Applications, 47(1): 7–11.

Sinha, G., Rani, R., and Dhir, R. (2012), “Handwritten Gurmukhi numeral recognition using zone-based hybrid feature extraction techniques”, International Journal of Computer Applications, 47(21):24–29.

Tang, M., Gao, H., Zhang, Y., Liu, Y., Zhang, P., and Wang, P. (2018), “Research on deep learning techniques in breaking text-based captchas and designing image-based captcha”, IEEE Transactions on Information Forensics and Security, 13(10): 2522-2537.

Thobhani, A., Gao, M., Hawbani, A., Ali, S.T.M., Abdussalam, A. (2020), “CAPTCHA Recognition Using Deep Learning with Attached Binary Images”, Electronics, 9(9): 1522. https://doi.org/10.3390/electronics9091522

Xu, Z., Qi, M., Wang, Z., Wen, S., Chen, S., and Xiang, Y. (2021). “IB2P: An imagebased privacy-preserving blockchain model for financial services”. IEEE International Conference on Blockchain (Blockchain). Dec 06-08. Melbourne, Australia.