IJE TRANSACTIONS B: Applications Vol. 30, No. 11 (November 2017) 1700-1706   

downloaded Downloaded: 118   viewed Viewed: 1422

H. Hassanpour, N. Samadiani and F. Akbarzadeh
( Received: July 31, 2017 – Accepted in Revised Form: September 08, 2017 )

Abstract    This paper proposes a new method to distinguish the printed digits, regardless of font and size, using neural networks.Unlike our proposed method, existing neural network based techniques are only able to recognize the trained fonts. These methods need a large database containing digits in various fonts. New fonts are often introduced to the public, which may not be truly recognized by the Optical Character Recognition (OCR). Therefore, the existing OCR systems may need to be retrained or their algorithm be updated. In this paper we propose a self-organizing map (SOM) neural network powered by appropriate features to achieve high accuracy rate for recognizing printed digits problem. In this method, we use a limited sample size for each digit in training step. Two expriments are designed to evaluate the performance of the proposed method. First, we used the method to classify a database including 2000 printed Persian samples with twenty different fonts and ten different sizes from which 98.05% accuracy was achieved. Second, the proposed method is applied to unseen data with different fonts and sizes with those used in training data set. The results show 98% accuracy in recognizing unseen data.


Keywords    Keywords: Recognition; multi-font; similarity measure; self-organizing map


چکیده    در این مقاله روشی جهت شناسایی مستقل از فونت و سایز ارقام چاپی فارسی توسط شبکه عصبی ارائه شده است. برخلاف روش ارائه شده این مقاله، روش­های موجود مبتنی بر شبکه عصبی، تنها قادر به شناسایی فونتهای آموزش دیده شده هستند. هر ساله فونتهای جدیدی به عموم معرفی می‌شود که این فونت‌ها نمی­توانند به درستی توسط سیستم‌های تشخیص الکترونیکی حروف شناسایی شوند. بنابراین، سیستم‌های موجود نیازمند بازآموزش برای به‌روزرسانی الگوریتم خود هستند. در این مقاله، روشی جهت شناسایی ارقام چاپی با دقت بالا توسط شبکه عصبی نگاشت خود سازمانده مجهز به ویژگی‌های مناسب ارائه داده­ایم. برای ارزیابی کارایی روش، دو آزمون طراحی شده است. در آزمون اول، داده­های مربوط به پایگاه داده­ای با ۲۰۰۰ نمونه ارقام چاپی فارسی حاوی ۱۲ فونت و ۱۰ سایز مختلف را دسته بندی نموده­ و به دقت ۰۵/۹۸%رسیده­ایم. در آزمون دوم روش ارائه شده برای شناسایی داده­های دیده نشده با فونت و سایز متفاوت از مجموعه آموزشی استفاده شده است. نتایج دقت ۹۸% برای شناسایی داده­های دیده نشده را نشان می­دهد.


1.      Sajedi, H., "Handwriting recognition of digits, signs, and numerical strings in persian", Computers & Electrical Engineering,  Vol. 49, (2016), 52-65.

2.      Al-Omari, S.A., Sumari, P., Al-Taweel, S.A. and Husain, A.J., "Digital recognition using neural network", Journal of Computer Science,  Vol. 5, No. 6, (2009), 427-434.

3.      Niu, X.-X. and Suen, C.Y., "A novel hybrid cnn–svm classifier for recognizing handwritten digits", Pattern Recognition,  Vol. 45, No. 4, (2012), 1318-1325.

4.      Man, Z., Lee, K., Wang, D., Cao, Z. and Khoo, S., "An optimal weight learning machine for handwritten digit image recognition", Signal Processing,  Vol. 93, No. 6, (2013), 1624-1638.

5.      Hanmandlu, M. and Murthy, O.R., "Fuzzy model based recognition of handwritten numerals", Pattern Recognition,  Vol. 40, No. 6, (2007), 1840-1854.

6.      Pradeep, J., Srinivasan, E. and Himavathi, S., "Neural network based recognition system integrating feature extraction and classification for english handwritten", International Journal of Engineering-Transactions B: Applications,  Vol. 25, No. 2, (2012), 99-106.

7.      Singh, M.P. and Dhaka, V., "Handwritten character recognition using modified gradient descent technique of neural networks and representation of conjugate descent for training patterns", Database,  Vol. 5, (2008), 145-158.

8.      Arora, S., Bhattacharjee, D., Nasipuri, M., Malik, L., Kundu, M. and Basu, D.K., "Performance comparison of svm and ann for handwritten devnagari character recognition", arXiv preprint arXiv:1006.5902,  (2010), 18-26.

9.      Shrivastava, V. and Sharma, N., "Artificial neural network based optical character recognition", arXiv preprint arXiv:1211.4385,  (2012), 73-80.

10.    Singh, D. and Khehra, B.S., "Digit recognition system using back propagation neural network", International Journal of Computer Science and Communication,  Vol. 2, No. 1, (2011), 197-205.

11.    El Kessab, B., Daoui, C., Bouikhalene, B., Fakir, M. and Moro, K., "Extraction method of handwritten digit recognition tested on the mnist database", International Journal of Advanced Science & Technology,  Vol. 50, (2013), 99-110.

12.    LeCun, Y., Bottou, L., Bengio, Y. and Haffner, P., "Gradient-based learning applied to document recognition", Proceedings of the IEEE,  Vol. 86, No. 11, (1998), 2278-2324.

13.    Mohebi, E. and Bagirov, A., "A convolutional recursive modified self organizing map for handwritten digits recognition", Neural Networks,  Vol. 60, (2014), 104-118.

14.    Boveiri, H.R., "Transformation-invariant classification of persian printed digits", International Journal of Signal Processing, Image Processing and Pattern Recognition,  Vol. 4, No. 3, (2011), 153-164.

15.    Karimi, H., Esfahanimehr, A., Mosleh, M., Salehpour, S. and Medhati, O., "Persian handwritten digit recognition using ensemble classifiers", Procedia Computer Science,  Vol. 73, (2015), 416-425.

16.    Ebrahimnezhad, H., Montazer, G.A. and Jafari, N., "Recognition of persian numeral fonts by combining the entropy minimized fuzzifier and fuzzy grammar", in Proceedings of the 6th Conference on 6th WSEAS Int. Conf. on Artificial Intelligence, Knowledge Engineering and Data Bases-Volume 6, World Scientific and Engineering Academy and Society (WSEAS)., (2007), 22-27.

17.    Montazer, G.A., Saremi, H.Q. and Khatibi, V., "A neuro-fuzzy inference engine for farsi numeral characters recognition", Expert Systems with Applications,  Vol. 37, No. 9, (2010), 6327-6337.

18.    Al-Abudi, B.Q., "Digit recognition using fractal and moment invariants", Iraq Journal of Science,  Vol. 50, No., (2009), 110-119.

19.    Dunlop, G., "A rapid computational method for improvements to nearest neighbour interpolation", Computers & Mathematics with Applications,  Vol. 6, No. 3, (1980), 349-353.

20.    Samadiani, N. and Hassanpour, H., "A neural network-based approach for recognizing multi-font printed english characters", Journal of Electrical Systems and Information Technology,  Vol. 2, No. 2, (2015), 207-218.

21.    Hassanpour, H., Darvishi, A. and Khalili, A., "A regression-based approach for measuring similarity in discrete signals", International Journal of Electronics,  Vol. 98, No. 9, (2011), 1141-1156.

22.    Kohonen, T. and Maps, S.-O., "Springer series in information sciences", Self-organizing maps,  Vol. 30, (1995).

23.    Tokunaga, K. and Furukawa, T., "Modular network som", Neural Networks,  Vol. 22, No. 1, (2009), 82-90.

24.             Lingras, P. and Butz, C.J., "Precision and recall in rough support vector machines", in Granular Computing, 2007. GRC 2007. IEEE International Conference on, IEEE. Vol., No. Issue, (2007), 654-654.

International Journal of Engineering
E-mail: office@ije.ir
Web Site: http://www.ije.ir