تعيين رفتار طبقه‌بندها با كليشه تصميم مبتني بر مدل مخفي ماركوف
الموضوعات :
هادی صدوقی یزدی
1
(دانشگاه تربیت معلم سبزوار)
الکلمات المفتاحية: رفتار طبقه‌بندهاشناسايي الگوكليشه تصميممدل مخفي ماركوفشبكه عصبي پس‌انتشار خطا,
ملخص المقالة :
مطالعه رفتار طبقه‌بندها از ديدگاه بررسي خطاهای آنها و ارائه راه حل مناسب براي كاهش خطا و افزايش كارآيي طبقه‌بندها مورد توجه است. عملكرد ضعيف سيستم شناسايي به دلايل تعداد كم نمونه‌هاي يادگير، نويز در داده‌ها، استفاده از ويژگي‌هاي شكننده به دليل عدم آگاهي كامل و تسلط كافي بر نوع الگو و استفاده از عملگرهای نامناسب در تعيين پاسخ سيستم، كاهش نرخ شناسايي در تصميم‌گيري نهايي را به دنبال دارد. با ارائه مدل مناسب آماري براي رفتار يا پاسخ يك سيستم شناسايي، مي‌توان عملكرد سيستم شناسايي را بهبود داد. در اين مقاله يك كليشه تصميم جديد كه با استفاده از مدل مخفي ماركوف ايجاد مي‌شود، رفتار نرون‌هاي يك شبكه عصبی پس‌انتشار خطا را، مدل مي‌كند. در روش‌هاي موجود ارتباط بين نرون‌ها و تأثير متقابل آنها در پاسخ به يك الگو مد نظر قرار نمي‌گيرد. ولي عملاً نرون‌هاي يك شبكه عصبي يا اجزاي يك طبقه‌بند، با هم بيان‌كننده عملكرد آن در قبال يك الگو هستند. بنابراين، با استفاده از يك كليشه تصميم جديد مبتني بر مدل مخفي ماركوف، ارتباط بين نرون‌هاي شبكه عصبي و نحوه پاسخ آن به نمونه‌هاي يادگير، مدل مي‌گردد تا از آن در شناسايي الگوهاي جديد استفاده شود. روش جديد در مدل‌كردن رفتار طبقه‌بند، در سه كاربرد شناسايي ارقام دست‌نويس فارسي، تشخيص ترافيك عادي در شبكه‌هاي اطلاع‌رساني و شناسايي نوع وسيله نقليه آزمون مي‌شود و افزايش قابل توجهي در نرخ شناسايي طبقه‌بند به دست مي‌دهد.
[1] R. Duda, P. Hart, and D. Stork, Pattern Classification, New York: John Wiley & Sons, 2001.
[2] H. Sameti, H. Sheikhzadeh, L. Deng, and R. L. Brennan, "HMMbased strategies for enhancement of speech signals embedded in nonstationary noise," IEEE Trans. on Speech and Audio Processing,vol. 6, no. 5, pp. 445-455, Sep. 1998.
[3] H. R. Kim and H. S. Lee, "Postprocessor using fuzzy vector quantiser in HMM - based speech recognition," Electronics Letters, vol. 27, no. 22, pp. 1998-2000, Oct. 1991.
[4] Y. Bengio, R. D. Mori, G. Flammia, and R. Kompe, "Global optimization of a neural network - hidden markov model hybrid," IEEE Trans. on Neural Networks, vol. 3, no. 2, pp. 252-259,Mar. 1992.
[5] C. Dugast, L. Devillers, and X. Aubert, "Combining TDNN and HMM in a hybrid system for improved continuous - speech recognition," IEEE Trans. on Speech and Audio Processing, vol. 2,no. 1, part II, pp. 217-223, Jan. 1994.
[6] C. A. Shipp and L. I. Kuncheva, "Relationship between combination methods and measures of diversity in combining classifiers," Information Fusion, vol. 3, no. 2, pp. 135-148, Jun. 2002.
[7] I. Bloch, "Information combination operators for data fusion: a comparative review with classification," IEEE Trans. on Systems, Man, and Cybernetics - Part A: Systems and Humans, vol. 26, no. 1, pp. 52-67, Jan. 1996.
[8] H. Bourlard and C. J. Wellekens, "Link between markov models and multilayer perceptrons," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 12, no. 12, pp. 1167-1178, Dec. 1990.
[9] Y. Bengio, R. Cardin, R. D. Mori, and Y. Normandi, "A hybrid coder for hidden markov models using a recurrent neural network," in Proc. IEEE Conf. Acoust., Speech, Signal Process., pp. 537-540,Albuquerque, NM, US, Apr. 1990.
[10] M. Franzini, K. F. Lee, and A. Waibel, "Connectionist Viterbi training: a new hybrid method for continuous speech recognition," in Proc. IEEE Conf. Acoust., Speech, Signal Process., pp. 425-428,Albuquerque, NM, US, Apr. 1990.
[11] L. T. Niles and H. F. Silverman, "Combining hidden markov models and neural network classifiers," in Proc. IEEE Conf. Acoust. Speech, Signal Process., pp. 417-420, Albuquerque, NM, US, Apr. 1990.
[12] S. J. Young, "Competitive training in hidden markov models," in Proc. IEEE Conf. Acoust. Speech, Signal Process., pp. 681-684, Albuquerque, NM, US, Apr. 1990.
[13] J. N. Hwang, J. A. Viontzos, and S. Y. Kung, "A systolic neural network architecture for hidden markov models," IEEE Trans. Acoust. Speech, Signal Process., vol. 37, no. 12, pp.1967-1979,Dec. 1989.
[14] J. S. Bridle, "Training stochastic model recognition algorithms as networks can lead to maximum, matual information of parameters," in Advances in Neural Information Processing Systems 2, pp. 211-217, 1990.
[15] S. Reiter, B. Schuller, and G. Rigoll, "A combined LSTM - RNN -HMM - approach for meeting event segmentation and recognition," in IEEE Int. Conf. on Acoustics, Speech and Signal Processing ,ICASSP 2006, vol. 2, pp. 14-19, May 2006.
[16] X. Wang and G. Dai, "A novel method to recognize complex dynamic gesture by combining HMM and FNN models," in Proc. IEEE Symposium on Computational Intelligence in Image and Signal Processing, CIISP 2007, pp. 13-18, Apr. 2007.
[17] B. Vitoantonio, D. Domenico, C. Lucia, and M. Giuseppe, "Pseudo 2D hidden markov models for face recognition using neural network coefficients," in Proc. IEEE Workshop on Automatic Identification Advanced Technologies, pp. 107-111, Jun. 2007.
[18] G. Zavaliagkos, Y. Zhao, R. Schwarts, and J. Makhoul, "A hybrid segmental neural net/hidden markov model system for continuous speech recognition," IEEE Trans. on Speech and Audio Processing,vol. 2, no. 1, pt. II, pp. 151-160, Jan. 1994.
[19] C. Dugast, L. Devillers, and X. Aubert, "Combining TDNN and HMM in a hybrid system for improved continuous - speech recognition," IEEE Trans. on Speech and Audio Processing, vol. 2, no. 1, part II, pp. 217-223, Jan. 1994.
[20] N. Morgan and H. Bourlard, "Continuous speech recognition using multilayer perceptrons with hidden markov models," in Proc. IEEE Conf. Acoust. Speech, Signal Process., pp. 413-416, Albuquerque,NM, US, Apr. 1990.
[21] Y. Bengio, R. D. Mori, G. Flammia, and R. Kompe, "Global optimization of a neural network - hidden markov model hybrid," IEEE Trans. on Neural Network, vol. 3, no. 2, pp. 252-259,Mar. 1992.
[22] G. Rigoll, "Maximum mutual information neural networks for hybrid connectionist - HMM speech recognition systems," IEEE Trans. on Speech and Audio Processing, vol. 2, no. 1, pt. II, pp. 175-184,Jan. 1994.
[23] M. Robinson, M. R. Azimi - Sadjadi, and J. Salazar, "Multi – aspect target discrimination using hidden markov models and neural networks," IEEE Trans. on Neural Networks, vol. 16, no. 2, pp. 447-459, Mar. 2005.
[24] S. B. Cho, "Neural - network classifiers for recognizing totally unconstrained handwritten numerals," IEEE Trans. on Neural Network, vol. 8, no. 1, pp. 43-53, Jan. 1997.
[25] J. Yang, Y. Xu, and C. S. Chen, "Hidden markov model approach to skill learning and its application to tele robotics," IEEE Trans. on Robotics and Automation, vol. 10, no. 5, pp. 621-631, Oct. 1994.
[26] N. Johnson, A. Galata, and D. Hogg, "The acquisition and use of interaction behavior models," in IEEE Computer Society Conf. on Computer Vision and Pattern Recognition, pp. 866-871, Jun. 1998.
[27] M. Brand and V. Kettnaker, "Discovery and segmentation of activities in video," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 22, no. 8, pp. 844-851, Aug. 2000.
[28 ] ها. صدوقي يزدي، م. لطفي زاد، ا. كبير و م. فتحي، "مدل تعاملي براي تعيين، عادي بودن رفتار انفرادي راننده،" نشريه فني و مهندسي مدرس، شماره 30 ، صص21-38 زمستان1386
[29] S. Vaseghi, Advanced Signal Processing and Digital Noise Reduction, John Wiley & Sons Ltd, 1996.
[30] C. Y. Suen, M. Berthod, and S. Mori, "Automatic recognition of handprinted characters the state of the art," in Proc. IEEE, vol. 68, no. 4, pp. 469-487, Apr. 1980.
[31] K. Badie and M. Shimura, "Machine recognition of Arabic cursive scripts," in Pattern Recognition in Practice, E. S. Gelsema and L. N. Kanal (eds.), pp. 315-323, North Holland Publishing Company,1980.
[32 ] و. جوهري مجد و س. م. رضوي، "بازشناسي فازي ارقام دس تنويس فارسي،" مجموعه مقالات اولين كنفرانس ماشين بينايي و پردازش تصوير ايران،صص151-144 ، دانشگاه بيرجند، اسفند 1379.
[33 ] ا. كبير و ر. عزمي، "ارائه دو الگوريتم براي شناسايي حروف چاپي فارسي،" مجموعه مقالات سومين كنفرانس سالانه بي نالمللي انجمن كامپيوتر ايران، صص 191-197 ، دانشگاه علم و صنعت ايران، دي. 1376
[34] K. Massruri and E. Kabir, "Recognition of hand - printed farsi characters by a fuzzy classifier," in Proc. of ACCV’95, vol. 2, pp. 607-610, Singapore, Dec. 1995.
[35 ] و. دست پاك و ر. صفابخش، "شناسايي حروف و علائم تايپي فارسي با استفاده از نماها،" مجموعه مقالات سومين كنفرانس مهندسي برق ايران، كامپيوتر، صص 198-206 ، دانشگاه علم و صنعت ايران، ارديبهشت
1374 [36 ] م. فهيمي و ر. ثاني، "تشخيص حروف دست نويس فارسي،" مجموعه مقالات اولين كنفرانس سالانه انجمن كامپيوتر ايران، ، صص110-103،دانشگاه صنعتی شریف،دی 1374.
[37 ] م. فتحي و ع. برومندنيا، "شناسايي ارقام و حروف مجزاي فارسي در شرايط نورغير يكنواخت،" مجموعه مقالات اولين كنفرانس سالانه انجمن كامپيوتر ايران،. 96صص 102-96 ، دانشگاه صنعتي شريف، دي 1374 .
[38 ] ك. فائز، ع. ختن زاد و م. شيرعلي شهرضا، "تشخيص حروف و ارقام دس تنويس فارسي با استفاده از گشتاورهاي شب هزرنيكي و به كمك شبكه هاي عصبي،"، صص240. -231 مجموعه مقالات سومين كنفرانس مهندسي برق ايران، دانشگاه علم و صنعت ايران، ارديبهشت 1374.
[39 ] ن. ساداتي و ب. نظري، "استفاده از منطق فازي در تشخيص ارقام دس تنويس - فارسي،" مجموعه مقالات سومين كنفرانس مهندسي برق ايران، صص. 247. 241 ، دانشگاه علم و صنعت ايران، ارديبهشت 1374.
[40] D. Anderson, T. F. Lunt, H. Javitz, A. Tamaru, and A. Valdes, Detecting Unusual Program Behavior Using the Statistical Component of the Next - generation Intrusion Detection Expert System (NIDES), Technical Report, SRI International, May 1995.
[41] H. Debar, M. Becker, and D. Siboni, "A neural network component for an intrusion detection system," in Proc. of the 1992 IEEE Computer Society Symposium on Research in Security and Privacy, pp. 240-250, Oakland, CA, May 1992.
[42] C. Dowell and P. Ramstedt, "The computer watch data reduction tool," in Proc. of the 13th National Computer Security Conf., pp. 99- 108, Washington, DC, Oct. 1990.
[43] L. T. Heberlein, G. Dias, K. Levitt, B. Mukherjee, J. Wood, and D. Wolber, "A network security monitor," in Proc. of the 1990 IEEE Symposium on Research in Security and Privacy, pp. 296-304,Oakland, CA, May 1990.
[44] S. R. Snapp, J. Brentano, G. Dias, T. Goan, L. T. Heberlein, C. Ho,K. Levitt, B. Mukherjee, S. Smaha, T. Grance, D. Teal, and D. Mansur, "DIDS (Distributed Intrusion Detection System) -motivation, architecture, and an early prototype," in Proc. of the 14th National Computer Security Conf., pp. 167-176, Washington, DC, US, Oct. 1991.
[45] P. G. Neumann and P. A. Porras, Experience with EMERALD to Date, First USENIX Workshop on Intrusion Detection and Network Monitoring, Santa Clara, CA, Apr. 1999.
[46] Information Exploration Shootout, URL: http://iris.cs.uml.edu:8080/network.html, 1998.
[47] V. Jacobson, C. Leres, and S. McCanne, Tcpdump, ftp://ftp.ee.lbl.gov, 1998.
[48] E. Prez " Image processing for intelligent transportation systems: application to road sign recognition," in Smart Imaging Systems, pp. 207-232, 1st ed., Bellingham, Washington: SPIE Press Monograph,2001.
[49] B. Bhanu, "Automatic target recognition: state of the art survey," IEEE Trans. Aerospace Electron System, vol. 22, no. 4, pp. 364-379, Jul. 1986.
[50] B. Dasarathy, "Information processing for target recognition from autonomous vehicles," in Proc SPIE Electro - Opt Tech Autonomous Vehicles, vol. 219, pp. 86-93, Feb. 1980.
[51] B. Bhanu and T. Jones, "Image understanding research for automatic target recognition," IEEE Aerospace Electron System Magazine, vol. 10, no. 8, pp. 15-22, Oct. 1993.
[52] M. A. Turk and A. P. Pentland, "Face recognition using eigenfaces,"in Proc. IEEE Conf. on Computer Vision and Pattern Recognition, pp. 586-591, Jun. 1991.
[53] D. Terrieri, "The eigenspace separation transform for neural network classifiers," Neural Networks, vol. 12, pp. 419-427, Apr. 1999.
