ادغام شبکه‌های عصبی بر اساس یادگیری با همبستگی منفی در بازشناسی برون‌خط کلمات دست‌نویس
محورهای موضوعی : مهندسی برق و کامپیوتر
سیدعلی‌اصغر عباس‌زاده آرانی
1
*
,
احسان‌اله کبیر
2
1 - دانشگاه تربیت مدرس
2 - دانشگاه تربیت مدرس
کلید واژه: ادغام طبقه‌بندها بازشناسی کلمات دست‌نویس پرسپترون چندلایه یادگیری با همبستگی منفی,
چکیده مقاله :
در این تحقیق، یک روش طبقه‌بندی جمعی بر اساس یادگیری با همبستگی منفی برای بازشناسی کل‌نگر کلمات دست‌نویس با حجم محدود پیشنهاد می‌شود. در این روش مجموعه داده آموزشی پس از پیش‌پردازش و استخراج ویژگی به طبقه‌بندهای پایه پرسپترون چندلایه اعمال می‌شود. سپس شبکه‌های عصبی پایه به روش یادگیری با همبستگی منفی، آموزش داده شده و از این طریق گوناگون می‌شوند. هنگامی که داده‌های آزمایشی پس از استخراج ویژگی به طبقه‌بندهای پایه اعمال می‌شوند، هر طبقه‌بند پایه خروجی نسبتاً متفاوتی را تولید می‌کند. با ادغام خروجی طبقه‌بندهای پایه، خروجی نهایی سیستم به دست می‌آید. برای آزمایش روش پیشنهادی از سه نوع ویژگی شامل ویژگی‌های مبتنی بر منطقه‌بندی، گرادیان تصویر و کد زنجیره‌ای کانتور استفاده شده است. در آزمایش‌هایی که روی 775 تصویر از نام 31 مرکز استان کشور، از مجموعه داده "ایرانشهر" انجام شده است، استفاده از ویژگی‌های مبتنی بر گرادیان و آموزش 6 شبکه پرسپترون با همبستگی منفی و ادغام آنها از طریق رأی‌گیری، میانگین نرخ بازشناسی برابر با 10/96 درصد را به دست داده است. سپس خطاهای روش پیشنهادی تحلیل و ردیابی شده است.
In this study, an ensemble classification method, based on negative correlation learning, is used for holistic recognition of handwritten words with limited vocabulary. In this method, training data set, after preprocessing and feature extraction, is applied to the base Multilayer Perceptron classifiers. These classifiers are trained by negative correlation learning to make them diverse. Features extracted from a test input are applied to the base classifiers, which produce somehow diverse outputs. By combining these outputs, the final output of the system is obtained. For experiments, three feature sets based on zoning, gradient image and contour chain code are extracted from the images. In experiments, performed on 775 images of 31 Province centers from "Iranshahr" dataset, when gradient-based features were used to train 6 Multilayer Perceptron classifiers by negative correlation, by Fusion the outputs of these classifiers through voting, an average recognition rate of 96.10 percent is achieved.
[1] A. Brakensiek and G. Rigoll, "Handwritten address recognition using hidden markov models," Reading and Learning, vol. 2956, pp. 103-122, 2004.
[2] R. Palacios, A. Gupta, and P. S. Wang, "Handwritten bank check recognition of courtesy amounts," Int'l J. of Image and Graphics, vol. 4, no. 2, pp. 203-222, Apr. 2004.
[3] M. Ye, P. A. Viola, S. Raghupathy, H. Sutanto, and C. Li, "Learning to group text lines and regions in freeform handwritten notes," in Proc. 9th Int. Conf. on Document Analysis and Recognition, IEEE Computer Society, vol. 1, pp. 28-32, Sep. 2007.
[4] V. Govindaraju and H. Xue, "Fast handwriting recognition for indexing historical documents," in Proc. of 1st Int'l Workshop on Document Image Analysis for Libraries, pp. 314-320, 2004.
[5] ک. مسروری، ا. کبیر و ک. بدیع، "شناسایی کلمات دست‌نویس فارسی با استفاده از الگوریتم DTW،" فصل‌نامه علمی پژوهشی دانشگاه شاهد، شماره 25، صص. 59-68،1378.
[6] M. Dehghan, K. Faez, M. Ahmadi, and M. Shridhar, " Handwritten Farsi (Arabic) word recognition: a holistic approach using discrete HMM," Pattern Recognition, vol. 34, no. 5, pp. 1057-1065, May 2001.
[7] M. Dehghan, K. Faez, M. Ahmadi, and M. Shridhar, " Unconstrained Farsi handwritten word recognition using fuzzy vector quantization and hidden Markov models," Pattern Recognition Letters, vol. 22, no. 2, pp. 209-214, Feb. 2001.
[8] A. Broumandnia, J. Shanbehzadehb, and M. Rezakhah, "Persian/Arabic handwritten word recognition using M-band packet wavelet transform," Image and Vision Computing, vol. 26, no. 6, pp. 829-842, Jun. 2008.
[9] G. Brown, Diversity in Neural Network Ensembles, Ph.D. Thesis, University of Birmingham, 2003.
[10] L. I. Kuncheva, Combining Pattern Classifiers: Methods and Algorithms, Wiley-Interscience, 2004.
[11] L. Xu, A. Krzyzak, and C. Y. Suen, "Methods of combining multiple classifiers and their applications to handwriting recognition," IEEE Trans. on Systems, Man, and Cybernetics, vol. 22, no. 3, pp. 418-435, May/Jun. 1992.
[12] B. Plessis, A. Sicsu, L. Heutte, E. Menu, E. Lecolinet, O. Debon, and J. V. Debon, "A multi-classifier combination strategy for the recognition of handwritten cursive words," in Proc. Int. Conf. on Document Analysis and Recognition, pp. 642-645, Oct. 1993.
[13] S. Gunter and H. Bunke, "Ensembles of classifiers for handwritten word recognition," International J. on Document Analysis and Recognition, vol. 5, no. 1, pp. 224-232, Jul. 2003.
[14] Y. S. Huang and C. Y. Suen, "A method of combining multiple experts for the recognition of unconstrained handwritten numerals," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 17, no.1, pp. 90-94, Jan. 1995.
[15] B. Verma, P. Gader, and W. Chen, "Fusion of multiple handwritten word recognition techniques," Patter Recognition Letters, vol. 22, no. 9, pp. 991-998, Jul. 2001.
[16] N. Farah, M. T. Khadir, and M. Sellami, "Artiffcial neural network fusion: application to Arabic words recognition," in Proc. European Symposium on Artificial Neural Network, pp. 151-156, Apr. 2005.
[17] R. Ebrahimpour, M. Amini, and F. Sharifizadehi, "Farsi handwritten recognition using combining neural networks based on stacked generalization," International J. on Electrical Engineering and Informatics, vol. 3, no. 2, pp. 146-164, Jun. 2011.
[18] H. El Abed and V. Margner, "A framework for the combination of different arabic handwritten word recognition systems," in Proc. 20th Int. Conf. on Pattern Recognition, pp. 1904-1907, Aug. 2010.
[19] Y. Kessentini, T. Burger, and T. Paquet, "Evidential combination of multiple HMM classifiers for multi-script handwritting recognition," Computational Intelligence for Knowledge-Based Systems Design, Lecture Notes in Computer Science, vol. 6178, pp. 445-454, 2010.
[20] T. Burger, Y. Kessentini, and T. Paquet, "Dempster-shafer based rejection strategy for handwritten word recognition," in Proc. Int. Conf. on Document Analysis and Recognition, pp. 528-532, Sep. 2011.
[21] R. Ebrahimpour, R. D. Vahid, and B. Mazloom Nezhad, "Decision templates with gradient based features for Farsi handwritten word recognition," International J. of Hybrid Information Technology, vol. 4, no. 1, pp. 1-12, Jan. 2011.
[22] H. Salimi and D. Giveki, "Farsi/Arabic handwritten digit recognition based on ensemble of SVD classifiers and reliable multi-phase PSO combination rule," International J. on Document Analysis and Recognition, vol. 16, no. 4, pp. 371-386, Dec. 2012.
[23] V. Frinken, T. Peter, A. Fischer, H. Bunke, T. M. T. Do, and T. Artieres, "Improved handwriting recognition by combining two forms of hidden markov models and a recurrent neural network," Computer Analysis of Images and Patterns, Lecture Notes in Computer Science, vol. 5702, pp. 189-196, 2009.
[24] Y. C. Chim, A. A. Kassim, and Y. Ibrahim, "Dual classifier system for hand printed alphanumeric character recognition," Pattern Analysis and Application, vol. 1, no. 3, pp. 155-162, Sep. 1998.
[25] R. M. O. Cruz, G. D. C. Cavalcanti, and T. I. Ren, "An ensemble classifier for offline cursive character recognition using multiple feature extraction techniques," in Proc. Int. Joint Conf. on Neural Networks, 8 pp., Jul. 2010.
[26] R. M. O. Cruz, G. D. C. Cavalcanti, and T. I. Ren, "Handwritten digit recognition using multiple feature extraction techniques and classifier ensemble," in Proc. 17th Int. Conf. on Systems, Signals, and Image Processing, pp. 215-218, 2010.
[27] M. Liwicki and H. Bunke, "Combining diverse on-line and off-line systems for handwritten text line recognition," Pattern Recognition, vol. 42, no. 12, pp. 3254-3263, Dec. 2009.
[28] M. Liwicki, H. Bunke, J. A. Pittman, and S. Knerr, "Combining diverse systems for handwritten text line recognition," Machine Vision and Applications, vol. 22, no. 1, pp. 39-51, Jan. 2011.
[29] C. Leila, K. Maamar, and C. Salim, "Combining neural networks for Arabic handwriting recognition," in Proc. 10th Int. Symp. on Programming and Systems, pp. 74-79, Apr. 2011.
[30] A. Rahman and B. Verma, "Effect of ensemble classifier composition on offline cursive character recognition," Information Processing and Management, vol. 49, no. 4, pp. 852-864, Jul. 2013.
[31] N. Azizi, N. Farah, and M. Sellami, "Ensemble classifier construction for Arabic handwritten recognition," in Proc. Int. Workshop on Systems, Signal Processing, and their Applications, pp. 272-274, May 2011.
[32] L. S. Oliveira, M. Morita, and R. Sabourin, "Feature selection for ensembles applied to handwriting recognition," International J. of Document Analysis and recognition, vol. 8, no. 4, pp. 262-279, Sep. 2006.
[33] B. E. Rosen, "Ensemble learning using decorrelated neural networks," Connection Science, vol. 8, no. 3-4, pp. 373-384, Jul. 1996.
[34] Y. Liu and X. Yao, "Ensemble learning via negative correlation," Neural Networks, vol. 12, no. 10, pp. 1399-1404, Dec. 1999.
[35] Y. Liu and X. Yao, "Simultaneous training of negatively correlated neural networks in an ensemble," IEEE Trans. on Systems, Man, and Cybernetics-Part B: Cybernetics, vol. 29, no. 6, pp. 716-725, Dec. 1999.
[36] C. L. Liu, K. Nakashima, H. Sako, and H. Fujisawa, "Handwritten digit recognition: investigation of normalization and feature extraction techniques," Pattern Recognition, vol. 37, no. 2, pp. 265-279, Feb. 2004.
[37] R. Azmi and E. Kabir, "A new segmentation technique for omnifont Farsi text," Pattern Recognition Letters, vol. 22, no. 2, pp. 97-104, Feb. 2001.
[38] Ø. D. Trier, A. K. Jain, and T. Taxt, "Feature extraction methods for character recognition-a survey," Pattern Recognition, vol. 29, no. 4, pp. 641-662, Apr. 1995.
[39] R. Polikar, "Ensemble based systems in decision making," Circuits and Systems Magazine, IEEE, vol. 6, no. 3, pp. 21-45, Third Quarter 2006.
[40] س. ح. نبوی کریزی و ا. کبیر، "تركيب طبقه‌بندها: ايجاد گوناگوني و قواعد تركيب،" نشريه علمي پژوهشي انجمن كامپيوتر ايران، جلد ۳، شماره ۳- الف، صص. 95-107، پایيز 1384.
[41] م. ع. باقری، غ. ع. منتظر و ا. کبیر، "سیستم‌های دسته‌بند چندگانه روش‌های طراحی و قواعد ترکیب شورا،" دو‌فصل‌نامه پردازش علایم و داده‌ها، جلد 8، شماره 2، صص. 29-56، 1390.
[42] L. I. Kuncheva and C. J. Whitaker, "Measures of diversity in classifier ensembles and their relationship with the ensemble accuracy," Machine Learning, vol. 51, no. 2, pp. 181-207, May 2003.
[43] D. Ruta and B. Gabrys, "Analysis of the correlation between majority voting error and the diversity measures in multiple classifier systems," in Proc. of the 4th Int. Symp. on Soft Computing, 7 pp., Paisley, UK, Jun. 2001.
[44] م. صنیعی آباده، س. محمودی و م. طاهرپور، داده‌کاوی کاربردی، نیاز دانش، صص. 187-189، چاپ اول 1391.