Responsive image
博碩士論文 etd-0811103-181724 詳細資訊
Title page for etd-0811103-181724
論文名稱
Title
雜訊環境下語音辨識系統之設計研究
A Design of Speech Recognition System under Noisy Environment
系所名稱
Department
畢業學年期
Year, semester
語文別
Language
學位類別
Degree
頁數
Number of pages
64
研究生
Author
指導教授
Advisor
召集委員
Convenor
口試委員
Advisory Committee
口試日期
Date of Exam
2003-07-25
繳交日期
Date of Submission
2003-08-11
關鍵字
Keywords
不特定語者、頻譜相減法、倒頻譜平均值消去、語詞辨識、隱藏式馬可夫模型
cepstral mean subtraction, spectral subtraction, speaker-independent, phrase recognition, hidden Markov model
統計
Statistics
本論文已被瀏覽 5658 次,被下載 0
The thesis/dissertation has been browsed 5658 times, has been downloaded 0 times.
中文摘要
本論文在於研究雜訊環境下的語音辨識系統,其目的在於現實吵雜環境中,能使語音辨識系統得以順利應用。首先,先以加強版頻譜相減法(Enhanced Spectral Subtraction,ESS)濾除雜訊聲音,經由梅爾倒頻譜平均消去法(MFCC with Cepstral Mean Subtraction)粹取語詞特徵,最後以隱藏式馬可夫模型(Hidden Markov Model,HMM)為基礎,建立一不特定語者(Speaker-Independent)在雜訊環境下之語詞辨識系統。

本研究所採用的語音資料庫為麥克風中文語料(股票上市公司名稱),再分別加上不同強度的噪音,作為評比依據,實現不特定語者(Speaker-Independent)在雜訊環境下之語詞辨識系統。
Abstract
The objective of this thesis is to build a phrase recognition system under noisy environment that can be used in real-life. In this system, the noisy speech is first filtered by the enhanced spectral subtraction method to reduce the noise level. Then the MFCC with cepstral mean subtraction is applied to extract the speech features. Finally, hidden Markov model (HMM) is used in the last stage to build the probabilistic model for each phrase.

A Mandarin microphone database of 514 company names that are in Taiwan’s stock market is collected. A speaker independent noisy phrase recognition system is then implemented. This system has been tested under various noise environments and different noise strengths.
目次 Table of Contents
致謝……………………………………………………………………Ⅰ
論文摘要………………………………………………………………Ⅱ
目錄……………………………………………………………………Ⅲ
圖表目錄………………………………………………………………Ⅵ

第一章 緒論............................................1
1-1 研究動機與目的......................................1
1-2 噪音環境下語辭辨識方法介紹..........................2
1-2-1 語音加強法......................................2
1-2-2 模型補償法......................................2

第二章 語詞辨識系統與數位語音訊號處理..............3
2-1 語詞辨識系統介紹....................................3
2-2 辨識系統之語音前置處理..............................5
2-2-1 端點偵測(Endpoint Detection) ......................5
2-2-2 乘上視窗函數(Window) ...........................7
2-3 語音訊號之特徵萃取.................................11
2-3-1倒頻譜(Cepstrum)................................12
2-3-2倒頻譜平均值消去法.............................15
2-3-3梅爾倒頻譜係數.................................16

第三章 語音訊號加強..................................20
3-1 頻譜相減法.........................................20
3-1-1 輸入輸出訊號...................................21
3-1-2 偏移量消去及半波整流...........................22
3-1-3 去除殘留雜訊...................................23
3-1-4 衰減無聲段訊號.................................24
3-2加強版頻譜相減法....................................26

第四章 隱藏式馬可夫模型為基礎之語詞辨識系統...... 29
4-1 語音訊號之隱藏式馬可夫模型......................29
4-2 隱藏式馬可夫模型之建立.............................30
4-3 隱藏式馬可夫模型之訓練.............................31
4-3-1 期望值最大演算法 ..............................31
4-3-2 參數重估(Parameters Reestimation)..............32
4-4 隱藏式馬可夫模型之辨識程序.........................37
第五章 系統設計實作結果與比較.......................40
5-1 系統設計...........................................40
5-2 雜訊介紹...........................................41
5-3 雜訊平均頻譜圖.....................................44
5-4 結果與比較.........................................47
5-4-1引擎室雜訊(Engine room noise)....................47
5-4-2 戰鬥機雜訊(F-16 cockpit noise)..................48
5-4-3 工廠雜訊(Factory noise)........................49
5-4-4 工廠雜訊(Factory noise)..........................50
5-4-5汽車雜訊(Car noise).............................51
5-4-6白高斯雜訊(White noise).........................52
5-4-7 吹風機雜訊(Blower noise).......................53

第六章 結論與建議.................................55
6-1 結論...............................................55
6-2 建議...............................................56

參考文獻...............................................57
附錄(一)...........................................59
參考文獻 References
[1]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,2002

[2]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,2002

[3]S.F. Boll,“Suppression of Acoustic Noise in Speech Using Spectral Subtraction.”IEEE Trans. On Acoust., Speech and Signal Proc., vol. ASSP-27, pp. 113-120, April 1979.

[4]P. Lockwood and J. Boudy, “Experiments with a Nonlinear Spectral Subtractor(NSS),Hidden Markov Models and the projection, for robust speech recognition in cars.”Speech Commun., vol. 11, pp. 215-228, June 1992.

[5]M.J.F Gales and S.J. Young,“Robust continuous speech recognition using parallel model combination.”IEEE Trans. Speech and Audio Proc., vol. 4, pp. 352-359, Sept. 1996

[6]J. R. Deller, J. G. Proakis and J. H. L. Hansen, Discrete Time Processing of Speech Signals, New York: Macmillan Pub. Co., 1993.

[7]A,M,Kondoz, Digital Speech coding, New York: John Wiley &Sons Inc., 1994.

[8]S.B. Davis, P. Mermelstein,“Comparison of Parametric Representation for Monosyllabic Word Recognition in Continuously Spoken Sentences.”IEEE Trans. On ASSP-28, pp.357-366, 1980.


[9]John R. Deller, John G. Prooakls and John H. Hansen, Discrete-Time Processing of Speech Signals, Maxwell Macmillan international.

[10]R.W. Schafer and J.D. Markel, Eds., Speech Analysis, New York: IEEE Press, 1979.

[11]Yoshua Bengio, “Markovian Models for Sequential Data.” Neural Computing Surveys 2,pp.129-162,1999.

[12]L.R. Rabiner, “A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition.” Proc. IEEE Trans, vol. 77 , pp.257 -286 , Feb. 1989.

[13]B.H. Juang and L.R. Rabiner, “Mixture Autoregressive Hidden Markov Models for Speech Signals.” IEEE Trans. Speech and Audio Processing,vol.33 ,pp 1404-1413, 1985.

[14]S.R. Quackenbush, T.P. Barnwell III, and M.A. Clements, Objective Measures of Speech Quality, Prentice Hall Advanced Reference Series, Englewood Cliffs, NJ, 1988.

[15]S. Furui,“Cepstral Analysis Technique for Automatic Speaker Verifi- cation.”IEEE Trans. Acoustics, Speech and Signal Processing, vol. 29, No. 2, pp. 254-272, 1981.

[16]C. Mokel, J. Monne and D. Jouvet“On-Line Adaptation of a Speech Recognizer to Variations in Telephone Line Conditions.”Eurospeech’93, pp. 1247-1250.

[17] 噪音雜訊資料庫http://spib.rice.edu/spib/data/signals/noise/
電子全文 Fulltext
本電子全文僅授權使用者為學術研究之目的,進行個人非營利性質之檢索、閱讀、列印。請遵守中華民國著作權法之相關規定,切勿任意重製、散佈、改作、轉貼、播送,以免觸法。
論文使用權限 Thesis access permission:校內校外均不公開 not available
開放時間 Available:
校內 Campus:永不公開 not available
校外 Off-campus:永不公開 not available

您的 IP(校外) 位址是 18.212.102.174
論文開放下載的時間是 校外不公開

Your IP address is 18.212.102.174
This thesis will be available to you on Indicate off-campus access is not available.

紙本論文 Printed copies
紙本論文的公開資訊在102學年度以後相對較為完整。如果需要查詢101學年度以前的紙本論文公開資訊,請聯繫圖資處紙本論文服務櫃台。如有不便之處敬請見諒。
開放時間 available 已公開 available

QR Code