Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/18756
標題: 利用共同向量法辨識中文母音及硬顎子音
Using the Method of Common Vector to Recognize Mandarin Vowel and Palatal Consonant
作者: 洪瑞騰
Hong, Rui-Teng
關鍵字: k最近鄰居法;k-nearest neighbor method;共同向量法;梅爾頻率倒頻譜係數;method of common vector;Mel-frequency cepstrum coefficient
出版社: 統計學研究所
引用: [1] 王小川 (2004),“語音訊號處理”。台北市:全華。 [2] 王國榮 (2000),“Visual Basic 6.0 實戰講座”。台北市:旗標。 [3] 吳明哲,黃世陽 (1998),“Visual Basic 6.0 中文版學習範本”。台北市:松崗。 [4] 陳宛余,李宗寶 (2011),“探討梅爾頻率倒頻譜係數之特徵擷取對國語子音之影響”。碩士論文,國立中興大學統計學研究所,台中。 [5] 陳佳妤,李宗寶 (2011),“探討梅爾頻率倒頻譜係數之特徵擷取對國語母音之影響”。碩士論文,國立中興大學統計學研究所,台中。 [6] 楊鎮光 (2002),“Visual Basic 與語音辨識” 台北市:松崗。 [7] 張國清,李宗寶 (2005),“用K-means之動態時間軸校正法於國語數字之語音辨識”。碩士論文,國立中興大學應用數學研究所,台中。 [8] 鍾靖爵,李宗寶 (2011),“利用共同向量法以及最佳梅爾頻率倒頻譜之特徵辨識特定語者之中文單音”。碩士論文,國立中興大學統計學研究所,台中。 [9] 羅璟義,李宗寶 (2009),“利用權重式共同向量法於中字彙之特定語者中文單音辨識”。碩士論文,國立中興大學應用數學研究所,台中。 [10] Cover, T. M. and Hart, P. E. (1967). “Nearest Neighbor Pattern Classification”, IEEE Trans. On Information Theory, vol. IT-13, No. 1, pp. 21-27. [11] Gulmezoglu, M. B., Dzhafarov, V. and Barkana, A. (1999). “A novel approach to isolated word recognition”, IEEE Trans. On Speech and Audio Processing, vol. 7, No. 6, pp. 620-627. [12] Gulmezogul M. B., V Dzhafarov, and A.Barkana, 2001 “The Common Vector Approach and Its Relation to Principal Component Analysis”, IEEE Trans. on Speech and Audio Processing, vol. 9, no. 6, pp. 655-622. [13] Keskin, M. Gulmezoglu, M. B. Parlaktuna, O. and Barkana, A. (1996), “Isolated word recognition by extracting personal differences”, in Proc. 6 th Int.Conf. Signal Processing Applications and Technology, Boston, MA, pp. 1989-1992. [14] M. Bilginer Gulmezo˘glu, Vakıf Dzhafarov, Mustafa Keskin, and Atalay Barkana,(1999) “A Novel Approach to Isolated Word Recognition”, IEEE Trans.on Speech and Audio Processing, vol.7,no.6,November. [15] Rabiner, L. R. and Sambur, M. R. (1975), “An algorithm for determining the endpoints of isolated utterances”, The Bell System Technique Ournal,Vol.54, pp. 297-315.
摘要: 
本篇論文主要是探討1391個中文單音國字對於其母音和舌面子音之研究。首先主要利用梅爾頻率倒頻譜係數和轉移梅爾頻率倒頻譜係數求出其特徵值,再利用K最近鄰居法和共同向量法去建構其語音模型來進行比對,觀察在不同的參數下如特徵值維度、子音音框的個數、音框擺盪值等組合去比較何種組合會有較好的辨識率。此次實驗的語音資料庫是由七位不同語者和本人所錄製的十組語音去做辨識,於第一部分利用K最近鄰居法辨識舌面子音下特徵值維度為20、30,音框擺盪值為1,子音音框數為16、18下會有較佳的辨識率;在第二部分利用共同向量法母音取樣點為512,特徵值維度為40,音框擺盪值為2的情況下會有較佳的辨識率,其最佳辨識率接近92%。

The aim of this paper is to discuss the 1391 mandarin consonant words recogniti-
on for their vowel and palatal consonant. To construct the words recognition system involves in two major features Mel-frequency cepstrum coefficient (MFCC) and transform Mel-frequency cepstrum coefficient(Delta -MFCC) are obtained, then using the method of common vector and k-nearest neighbor (KNN) method to construct model. There are many factors may influence the rate of recognition such as the dimension of MFCC, the swing of frame, the length of frame and so on. The speech database in this experiment are recorded by eight speakers. Each isolated mandarin word is recorded ten times. The KNN is used for the first part, and the method of common vector for the second. Through the first part’s optimal parameters, the highest recognition rate we get is about 92% from second part.
URI: http://hdl.handle.net/11455/18756
其他識別: U0005-3006201217160600
Appears in Collections:統計學研究所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.