Abstract Persian handwritten digit recognition is one of the important topics of image processing which significantly considered by researchers due to its many applications. The most important challenges in Persian handwritten digit recognition is the existence of various patterns in Persian digit writing that makes the feature extraction step to be more complicated.Since the handcraft feature extraction methods are complicated processes and their performance level are not stable, most of the recent studies have concentrated on proposing a suitable method for automatic feature extraction. In this paper, an automatic method based on machine learning is proposed for high-level feature extraction from Persian digit images by using Convolutional Neural Network (CNN). After that, a non-linear multi-class Support Vector Machine (SVM) classifier is used for data classification instead of fully connected layer in final layer of CNN. The proposed method has been applied to HODA dataset and obtained 99.56% of recognition rate. Experimental results are comparable with previous state-of-the-art methods.

References

[1] Alizadehashraf B. and Roohi S., “Persian Handwritten Character Recognition Using Convolutional Neural Network,” in Proceedings of 10th Iranian Conference on Machine Vision and Image Processing, Isfahan, pp. 247-251, 2017.

[2] Boukharouba A. and Bennia A., “Novel Feature Extraction Technique for The Recognition of Handwritten Digits,” Applied Computing and Informatics, vol. 13, no. 1, pp. 19-26, 8 2017.

[3] Broumandnia A., Shanbehzadeh J., and Varnoosfaderani M., “Persian/Arabic Handwritten Word Recognition Using M-Band Packet Wavelet Transform,” Image and Vision Computing, vol. 26, no. 6, pp. 829-842, 2008.

[4] Chherawala Y. and Cheriet M., “Arabic Word Descriptor for Handwritten Word Indexing and Lexicon Reduction,” Pattern Recognition, vol. 47, no. 10, pp. 3477-3486, 2014.

[5] Chollet F., Deep Learning With Python, Manning Publications Co, 2018.

[6] Dehghan M., Faez K., Ahmadi M., and Shridhar M., “Handwritten Farsi (Arabic) Word Recognition: A Holistic Approach Using Discrete HMM,” Pattern Recognition, vol. 34, no. 5, pp. 1057-1065, 2001.

[7] Farahbakhsh E., Kozegar E., and Soryani M., “Improving Persian Digit Recognition By Combining Data Augmentation and Alexnet,” in Proceedings of 10th Iranian Conference on Machine Vision and Image Processing, Isfahan, pp. 265-270, 2017.

[8] Goodfellow I., Bengio Y., and Courville A., Deep Learning, MIT Press, 2016.

[9] Hajizadeh R., Aghagolzadeh A., and Ezoji M., “Manifold based Persian Digit Recognition Using the modified Locally Linear Embedding and Linear Discriminative Analysis,” in Proceedings of 2nd International Conference on Knowledge- Based Engineering and Innovation, Tehran, pp. 614-618, 2015.

[10] Ioffe S. and Szegedy C., “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift,” in Proceedings of the 32nd International Conference on Machine Learning, Lille, 448-456, 2015.

[11] Khorashadizadeh S. and Latif A., “Arabic/Farsi Handwritten Digit Recognition usin Histogra of Oriented Gradient and Chain Code Histogram,” The International Arab Journal of Information Technology, vol. 13, no. 4, pp. 367-374, 19 2016.

[12] Khosravi H. and Kabir E., “Introducing a Very Large Dataset of Handwritten Farsi Digits and A Study on Their Varieties,” Pattern Recognition Letters, vol. 28, no. 10, pp. 1133-1141, 2007.

[13] Kim P., “Matlab DEEP LEARNING,” With Machine Learning, Neural Networks and Artificial Intelligence, vol. 130, 2017.

[14] Liang H., Sun X., Sun Y., and Gao Y., “Text Feature Extraction Based on Deep Learning: A Review,” EURASIP Journal on Wireless Communications and Networking, vol. 2017, no. 1, pp. 1-12, 2017.

[15] Mozaffari S., Faez K., Märgner V., and El-Abed H., “Lexicon Reduction Using Dots for Off-Line Farsi/Arabic Handwritten Word Recognition,” Pattern Recognition Letters, vol. 29, no. 6, pp. 724-734, 2008.

[16] Naz S., Umar A., Ahmad R., Siddiqi I., Ahmed S. B., Razzak M., and Shafait F., “Urdu Nastaliq Recognition Using Convolutional-Recursive Deep Learning,” Neurocomputing, vol. 243, pp. 80-87, 13 2017.

[17] Nooraliei A., “Persian Handwritten Digits Recognition By Using Zoning and Histogram Projection,” in Proceedings 3rd Joint Conference of AI and Robotics and 5th RoboCup Iran Open International Symposium, Tehran, pp. 1-5, 2013.

[18] Parseh M. and Meftahi M., “A New Combined Feature Extraction Method for Persian Handwritten Digit Recognition,” International Journal of Image and Graphics, vol. 17, no. 02, pp. 1750012, 2017.

[19] Patterson J. and Gibson A., Deep learning: A Practitioner's Approach. O'Reilly Media, Inc, 2017.

[20] Pourmohammad A. and Ahadi S., “Using Single-Layer Neural Network for Recognition of Isolated Handwritten Persian Digits,” in Proceedings of 7th International Conference on Information, Communications and Signal Processing, Macau, pp. 1-4, 2009.

[21] Rashnodi O., Sajedi H., and Abadeh M., “Persian Handwritten Digit Recognition Using Support Vector Machines,” International Journal of Computer Applications, vol. 29, no. 12, pp. 1-6, 2011.

[22] Rashnodi O., Sajedi H., and Abadeh M., “Using Box Approach in Persian Handwritten Digits Recognition,” International Journal of Computer Applications, vol. 32, no. 3, pp. 1-8, 2011.

[23] Razzak M., Anwar F., Husain S., Belaid A., and Sher M., “HMM And Fuzzy Logic: A Hybrid Approach for Online Urdu Script-Based Languages’ Character Recognition,” Knowledge- Based Systems, vol. 23, no. 8, pp. 914-923, 2010.

[24] Saeed K. and Albakoor M., “Region Growing Based Segmentation Algorithm for Typewritten and Handwritten Text Recognition,” Applied Soft Computing, vol. 9, no. 2, pp. 608-617, 2009. 578 The International Arab Journal of Information Technology, Vol. 17, No. 4, July 2020

[25] Vapnik V., The Nature of Statistical Learning Theory, Springer science and Business Media, 2013.

[26] Zamani Y., Souri Y., Rashidi H., and Kasaei S., “Persian Handwritten Digit Recognition By Random Forest and Convolutional Neural Networks,” in Proceedings of 9th Iranian Conference on Machine Vision and Image Processing, Tehran, pp. 37-40, 2015. Mohammad Parseh received his first degree in Software Engineering from Shahid Chamran University of Ahwaz in 2008 and his MSc degree in Artificial Intelligence from Tabriz University in 2012. Currently, he is PhD candidate in Semnan University, Iran. His interested topics are handwritten character recognition, scene understanding, visual object tracking, deep learning, big data analysis and machine learning. Mohammad Rahmanimanesh received his MS and PhD both from the TarbiatModares University, Tehran, Iran, and BS from the Sharif University ofTechnology, Tehran, Iran, all in Computer Engineering. He is currently an Assistant Professor at Semnan University, Semnan, Iran. He is a member of IEEE and his research interests include network security, fuzzy systems, softcomputing, and data mining. Parviz Keshavarzi received the M.S. degree in electronic engineering from Tehran University, Tehran, Iran, in 1988 and the Ph.D. degree in electrical engineering from the University of Manchester, Manchester, U.K., in 1999. He is currently an Associate Professor with Semnan University, Semnan, Iran. His research interests include neuromorphics and nanoelectronics.