本論文探討如何利用梅爾倒頻譜參數、離散型隱藏式馬可夫模型及維特比演算法搭配時間動態校準等語詞辨識相關技術，來設計一套完整的中文關鍵語詞搜尋系統。在論文中，我們參考了各種對話系統的原理，實作了一個以對話方式來作輸入的高雄榮總自然語言掛號系統。系統在詢問使用者之身分證字號來完成身份確認後，使用者可在一句話內，完成門診掛號、查詢門診醫師、取消掛號等相關動作。平均每次掛號可在75秒內，完成所指定的動作。在實驗室中，語者獨立的環境下，可以達到97%的辨識率。

A Mandarin keyword spotting system based on MFCC, discrete-time HMM and Viterbi algorithm with DTW is proposed in this thesis. Joining with a dialogue system, this keyword spotting platform is further refined to a prototype of natural speech patient registration system of Kaohsiung Veterans General Hospital. After the ID number is asked by the computer-dialogue attendant in the registration process, the user can finish all relevant works in one sentence. Functions of searching clinical doctors, making and canceling registration are all built in this system. In a laboratory environment, the correct rate of this speaker-independent patient registration system can reach 97% and all registration process can be completed within 75 seconds.

目 錄
頁 次
致謝辭　………………………………………………………… I
論文摘要　……………………………………………………… II
目錄　…………………………………………………………… III
圖表目錄　………………………..………………………………… VI

第一章 緒論　…………………………………………………… 1
1-1 研究背景　……………………………………………… 1
1-2 研究目的　………………………………………………... 2
1-3 章節概要　………………………………………………... 3

第二章 語詞辨識系統與數位語音訊號處理　……………. 4
2-1 語詞辨識的基本架構　………………………………… 5
2-2 語音訊號的前置處理　………………………………… 6
2-2-1視窗函數　……………………………………………… 6
2-3 特徵萃取　……………………………………………… 12
2-3-1 倒頻譜係數　……………………………………… 12
2-3-2 梅爾倒頻譜係數　………………………………… 15
2-3-3 CMS（Cepstral Mean Subtraction）　………………. 18
2-4 Dynamic Time Warping與語詞辨識　………………… 20
2-4-1 Linear Time Warping　…………………………… 22
2-4-2 Dynamic Time Warping　………………………… 23
2-5 隱藏式馬可夫模型與語詞辨識　………………………. 28
2-5-1 模型描述　………………………………………… 28
2-5-2 參數初始化　……………………………………….. 28
2-5-3 隱藏式馬可夫模型之訓練　……………………… 29
2-5-4 期望值最大演算法（EM）　……………………….. 32
2-5-5 參數重估　………………………………………….. 33
2-5-6 隱藏式馬可夫模型之辨識程序　………………… 35

第三章 中文關鍵語詞搜尋法　…………………………… 38
3-1 以DTW為基礎的關鍵語詞搜尋法　………………….. 38
3-2 以HMM為基礎的關鍵語詞搜尋法　………………… 43

第四章 實驗設計、數據及討論　…………………………… 49
實驗一 以垃圾模型和反詞模型為基礎的關鍵語詞搜尋　.. 49
實驗二 高雄榮總自然語言掛號系統　…………………….. 51

第五章 結論與展望　…………………………………………. 59
5-1 結論　………………………………..…………………... 59
5-2 建議　………………………………….………………… 60

參考文獻　………………………………………………………….. 61
附錄　………………………………………………………………... 63

[1] A,M,Kondoz, Digital Speech coding, New York: John Wiley &Sons Inc., 1994.
[2] R.W. Schafer and J.D. Markel, Eds., Speech Analysis, New York: IEEE Press, 1979.
[3] J. R. Deller, J. G. Proakis, and J. H. L. Hansen, Discrete Time Processing of Speech Signals, New York: Macmillan Pub. Co., 1993.
[4] John R. Deller, John G. Prooakls, John H. Hansen, “Discrete-Time Processing Of Speech Signals”, Maxwell Macmillan international
[5] S. B. Davis, P. Mermelstein, “Comparison Of Parametric Representations For Monosyllabic Word Recognition In Continuously Spoken Sentences”, IEEE Transations on ASSP-28, pp 357-366, 1980
[6] S.S.Stevens and J. Volkmann, “The relation of pitch of frequency: Arevised scale,” Am.J.Psychol., 53:329-353,1940.
[7] 下萩原 勉, 浜田 玲子, 井手 一郎, 坂井 修一, 田中 英彦「テンプレート切出しによる不特定話者対応のワードスポッティング」第60回情処全国大会 (拓殖大学八王子校舎) No.2L-1, Vol.2, pp.131-132 (Mar 2000)
[8] Brown, M.; Rabiner, L.; “Dynamic time warping for isolated word recognition based on ordered graph searching techniques,” Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP '82. , Volume: 7 , May 1982 Page(s): 1255 -1258
[9] Binfeng Yan; Rui Guo; Xiaoyan Zhu; Bo Zhang; “An approach of keyword spotting based on HMM ,“ Intelligent Control and Automation, 2000. Proceedings of the 3rd World Congress on , Volume: 4 , 2000 ,Page(s): 2757 -2759 vol.4
[10] Wilpon, J.G.; Rabiner, L.R.; Lee, C.-H.; Goldman, E.R.; “Automatic recognition of keywords in unconstrained speech using hidden Markov models,” Acoustics, Speech, and Signal Processing, IEEE Transactions on , Volume: 38 Issue: 11 , Nov 1990 ,Page(s): 1870 -1878
[11] Eng-Fong Huang; Hsiao-Chuan Wang; Soong, F.K.; “A fast algorithm for large vocabulary keyword spotting application,” Speech and Audio Processing, IEEE Transactions on , Volume: 2 Issue: 3 , Jul 1994, Page(s): 449 -452
[12] Manhung Siu; Ostendorf, M.;” Variable n-grams and extensions for conversational speech language modeling,” Speech and Audio Processing, IEEE Transactions on , Volume: 8 Issue: 1 , Jan 2000 ,Page(s): 63 -75.
[13] Lawrence Rabiner; Biing-Hwang Juang, “Fundamentals of Speech Recognition”
[14] 楊鎮光, “Visual Basic與語音辨識—讓電腦聽話”, 民國91年6月17日.
[15] Michael Seltzer, “SPHINX III Signal Processing Front End Specification,” 31 August 1999.
[16] Shotaro Akaho, “Mixture model for image understanding and the EM algorithm,’ 7 April 1995.
[17] Jeff A. Bilmes, “A Gentle Tutorial of EM Algorithm and its Application Parameter Estimation for Gaussian Mixture and Hidden Markov Models”.
[18] 王怡理,“中文關鍵語詞搜尋系統之設計研究,” 國立中山大學電機工程研究所碩士論文, 民國92年1月8日.
[19] 賴昭華, “不特定語者中量語詞辨識系統之設計研究,” 國立中山大學電機工程研究所碩士論文, 民國91年9月12日.
[20] 鄭順德, “不特定語句中量語者辨識系統之設計研究,”國立中山大學電機工程研究所碩士論文, 民國91年9月13日.