臺灣博碩士論文加值系統

本研究使用SOCMAC類神經網路設計肺音分類器。為了尋找最適合萃取肺音特徵的小波基底，研究使用12種小波基底，包含了Coiflet-2、Coiflet-3、Coiflet-4、Coiflet-5、Daubechies-5、Daubechies-6、Daubechies-7、Daubechies-8、Symlet-5、Symlet-6、Symlet-7、Symlet-8。對於各種基底，我們調整分類器的參數，使其能達到最佳效果。實驗結果顯示，Symlet-8基底具有最佳的分類正確率。在最佳的結果參數下，分類器的學習正確率可達88.91%，而測試正確率可達到86.70%。

In this study, we use wavelet transform and SOCMAC neural networks to design lung sound classifiers. To find the best representation for lung sounds, we use 12 different wavelet bases to decomposition the lung sounds, which include Coiflet-2, Coiflet-3, Coiflet-4, Coiflet-5, Daubechies-5, Daubechies-6, Daubechies-7, Daubechies-8, Symlet-5, Symlet-6, Symlet-7, and Symlet-8. For every basis, the parameters of the classifier are selected so that the classifier has the best classification accuracy. Experiment results show that Symlet-8 has the best classification accuracy, out of the 12 different wavelet bases.

目錄
摘　要 i
ABSTRACT ii
誌　謝 iii
目錄 iv
表目錄 vi
圖目錄 viii
第一章 緒論 1
1.1 研究背景 1
1.2 文獻回顧 2
1.3 論文架構 7
第二章 肺音介紹 8
2.1 肺音的產生 8
2.2 支氣管音 8
2.3 肺泡音 9
2.4 氣管音 10
2.5 哮喘音 10
2.6 乾囉音 11
2.7 哮鳴音 12
2.8 爆裂音 13
2.9 嘎嘎聲 14
2.10 肺部羊鳴音 15
第三章 背景知識 16
3.1 小波轉換 16
3.2 SOCMAC網路 19
3.2.1 傳統的CMAC 19
3.2.2 自組織特徵圖 21
3.2.3 灰色關聯分析 22
3.2.4 SOCMAC網路的性質 23
3.2.4.1 核心概念 23
3.2.4.2 灰色信任度指配 24
3.2.4.3 記憶體大小 25
3.2.4.4 鄰近區域範圍 26
第四章 肺音前處理 28
4.1 肺音資料庫 28
4.2 資料處理 29
4.3 特徵值萃取 31
第五章 肺音分類器與分類結果 34
5.1 分類器設計 34
5.1.1 參數設定 35
5.1.2 終止條件 35
5.2 分類結果 36
第六章 結論 55
參考文獻 56

[1]T. Kaisla, A. Sovijarvi, P. Piirila, H-M. Rjala, S. Haltsonen, and T. Rosqvist. “Validated method for automatic detection of lung sound crackles”, Med. Biol Eng.Comput., Vol. 29, pp. 517-521(1991)
[2]B. Sankur, Y. P. Kahya, E. C. Guler, and T. Engin. “Comparison of AR-based algorithms for respiratory sounds classification”, Comput. Biol. Med., Vol. 24, pp.67-76(1996)
[3]L. J. Hadjileontiadis, and S. M. Panas. “Separation of discontinuous adventitious sounds from vesicular using a wavelet-based filter” IEEE Trans. Eng., Vol. 44 pp. 1269-1281(1997)
[4]L. Vannuccini, M. Rossi, and G.. Pasquali. “A new method to detect crackles in respiratory sounds” Tchnology and Health care pp.75-79(1998)
[5]Y. Shabtai-Musih, B. G. James, G.. Noam. “Spectral content of forced expiratory wheezes during air, He, and SF6 breathing in normal humans” Journal of Applied Physiology, vol. 72, no. 2, pp. 629-635(1992)
[6]Kevin E. Forkheim, D. Scuse, and H. Pasterkamp. “A comparison of neural network models for wheeze detection” IEEE Conference pp. 214-219(1995)
[7]Homs-Corbera, J. A. Fiz, J. Morera, and R. Jane. “Time-frequency detection and analysis of wheezes during forced exhalation” IEEE Trans. Biomed Eng., Vol. 51 pp. 182-186(2000)
[8]S. Pittner and S. V. Kamarthi. “Feature extraction from wavelet coefficients for pattern recongnition tasks” IEEE Transaction on pattern analysis and machine intelligence pp.83-88(1999)
[9]Kandaswamy, C. S. Kumar, R. P. Ramanathan, S. Jayaraman, and N. Malmurugan. “Neural classification of lung sound using wavelet coefficients” Comput. Biol Ned. pp. 523-537(2004)
[10]M. F. Yeh, K. C. Chang, “A self-organizing CMAC network with gray credit assignment”, IEEE Trans. Syst., Man, Cybern. B, Cybern, Vol. 36, no. 3, pp.623-635(2006)
[11]J. S. Albus, “Data storage in the crerbellar model articulation controller (CMAC)”, J. Dyn. Syst., Meas., Contr., Trans. ASME, vol. 97, 99. 228-233(1975)
[12]J. S. Albus, Brains, Behavior and Robotics, New York: McGraw-Hill(1981)
[13]W. T. Miller and F. H. Glanz, “CMAC: An associative of a bipedal walking alternative to backpropagation”, Proc. IEEE, vol. 78, pp. 1561-1567(1990)
[14]F. H. Glanz, E. T. Miller, amd L. G. Kraft, “An overview of the CMAC neural network”, Proc. IEEE Neural Network Ocean Eng., Washington, DC, pp. 301-308(1991)
[15]H. Kim and C. S. Lin, “Use of adaptive resolution for better CMAC learning”, Int. Joint Conf. Neural Network, vol. 1, Baltimore, MD, pp. 517-522(1992)
[16]S. Haykin, Neural Network: A Comprehensive Foundation, New York: Macmillan College Publishing Company (1994)
[17]M. C. Su, N. Declaris, and T. K. Liu, “Application of neural network in cluster analysis”, Proc. IEEE Int. Conf. Syst., Man and Cybern., Orlando, FL, pp. 1-6(1997)
[18]Rafael C. Gonzalez and Richard E. Woods, Digital image processing 2/E. Prentice Hall. (2002)
[19]Steven Lehrer, Understanding Lung Sounds Saunders.(1984)
[20]陳冠宏，多種肺音自動辨識之研究，碩士論文，國立台灣大學生物產業電機工程學研究所，台北(1996)