Naskh and Nastalique text recognition are a challenging task in the Pattern Recognition field because of the cursive and context sensitive nature of the script. Many languages use Naskh or/and Nastalique style for writing. Due to the complexities associated with these writing styles, not much effort has been done for the development of real-time recognition systems for Naskh and Nastalique writing style languages. Traditional recognition process segments the text image into characters for subsequent OCR phases which is less accurate for Naskh/Nastalique text and reduces the accuracy of the recognition system. Recently, Recurrent Neural Network (RNN) based Long Short Term Memory (LSTM) architecture with Connectionist Temporal Classification (CTC) has shown a remarkable result in text image recognition. This paper presents the recognition challenges in the Naskh and Nastalique writing style text and a study of different deep learning techniques applied for the recognition of Naskh Arabic and Nastalique Urdu text.

Naskh, Nastalique, Recognition Challenges, RNN, LSTM

[1] G.S. Lehal, “Choice of recognizable units for urdu OCR”, In Proceedings of the Workshop on Document Analysis and Recognition (DAR`12), pp.79–85, 2012.
[2] S. Garg, A.P. S, K. C, “An Extensive Survey on Text Detection and Recognition”, International Journal of Computer Sciences and Engineering, Vol.7, No.1, pp.546-551, 2019.
[3] A. Ul-Hasan, S.B. Ahmed, F. Shafait, T.M. Breuel, “Offine printed Urdu Nastaleeq script recognition with bidirectional LSTM networks”, In Proc. 12th Int. Conf. Document Analysis Recognition (ICDAR), pp.1061-1065, 2013.
[4] A. Ul-Hasan, M.Z. Afzal, F. Shafait, M. Liwicki, T.M. Breuel, “A sequence learning approach for multiple script identification”, In Document Analysis and Recognition (ICDAR),pp.1046-1050, 2015.
[5] A. Graves, S. Fern´andez, F. Gomez, and J. Schmidhuber, “Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks”, In Proceedings of the 23rd International Conference on Machine learning, pp.369–376, 2006.
[6] S.B. Ahmed, S. Naz, M.I. Razzak, S.F. Rashid, M.Z. Afzal, T.M. Breuel, “Evaluation of cursive and non-cursive scripts using recurrent neural networks”, Neural Comput. Appl. vol. 27, no. 3, pp.603-613, 2016.
[7] R. Ahmad, Z. Afzal, S.F. Rashid, “KPTI: Katib`s Pashto Text Imagebase and Deep Learning Benchmark”, 15th International Conference on Frontiers in Handwriting Recognition (ICFHR), 2016.
[8] A. Graves, M. Liwicki, S. Fernandez, R. Bertolami, H. Bunke, J. Schmidhuber, “A Novel Connectionist System for Unconstrained Handwriting Recognition”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, No. 5, pp.855–868,2009.
[9] S.F. Rashid, M. Schambach, J. Rottland, S. Nll, “Low resolution arabic recognition with multidimensional recurrent neural networks”, In Proceedings of the 4th International Workshop on Multilingual OCR, 2013.
[10] R. Ahmad, Z. Afzal, S.F. Rashid, M. Liwicki, T.M. Breuel, “Scale and rotation invariant ocr for Pashto cursive script using mdlstm network”, In Document Analysis and Recognition (ICDAR), pp.1101-1105, 2015.
[11] R. Ahmad, S. Naz, Z. Afzal, S.F. Rashid, M. Liwicki, A. Dengel, “Deepkhatt: A deep learning benchmark on arabic script”, Document Analysis and Recognition (ICDAR). 14th International Conference on IEEE, 2017.
[12] R. Maalej, N. Tagougui, and K. Kherallah, “Recognition of Handwritten Arabic Words with Dropout Applied in MDLSTM” ICIAR, pp.746–752, 2016.
[13] S. Naz, A. I. Umar, R. Ahmed, M.I. Razzak, S.F. Rashid, F. Shafait, “Urdu Nasta`liq text recognition using implicit segmentation based on multi-dimensional long short term memory neural networks”, Springer-Plus, vol. 5, no. 1, pp.1-16, 2016.
[14] S. Naz, A.I. Umar, R. Ahmad, S.B. Ahmed, S.H. Shirazi, I. Siddiqi, and M.I. Razzak, “Offline cursive Urdu-Nastaliq script recognition using multidimensional recurrent neural networks” Neurocomputing, vol. 177, pp. 228-241, 2016.
[15] S. Naz, A. I. Umar, R. Ahmad, S.B. Ahmed, S.H. Shirazi, M.I. Razzak, “Urdu Nasta`liq text recognition system based on multidimensional recurrent neural network and statistical features”, Neural Comput. Appl. vol. 28, no. 2, pp. 219-231, 2016.
[16] S. Naz, S.B. Ahmed, R. Ahmad, M.I. Razzak, “Zoning features and 2DLSTM for Urdu text-line recognition”, Procedia Computer Science Vol. 96, No. 1, pp.16-22, 2016.
[17] S. Naz, A. I. Umar, R. Ahmad, I. Siddiqi, S.B. Ahmed, M.I. Razzak, F. Shafait, “Urdu Nastaliq recognition using convolutional recursive deep learning” Neurocomputing, vol. 243, pp. 80-87, 2017.