Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/17884
標題: 利用 Discriminative Common Vectors 於國語數字之語音辨識
Using the Method of Discriminative Common Vectors to Speech Recognition of Mandarin digits
作者: 林凱昇
Lin, Kai-Sheng
關鍵字: Discriminative Common Vector;共同向量;特徵值
出版社: 應用數學系所
引用: [1]M. Bilginer Gulmezoglu, Vakif Dzhafarov, and Atalay Barkana ,“The common vector approach and its relation to principal component analysis” IEEE Trans. On Speech and Audio Processing, vol. 9. No. 6, 1999 [2]M. Keskin, M. B. Gulmezoglu, O. Parlaktuna, and A. Barkana, “Isolated word recognition by extracting personal differences,” in Proc. 6 th Int.Conf. Signal Processing Applications and Technology, Boston, MA , pp.1989-1992, 1996. [3]S. Yucel, “Application of Gram-Schmidt orthogonalization method to speech recognition for different noise levels” graduation project, Elect. Electron. Eng. Dept., Osmangazi Univ., Eskisehir, Turkey, 1996. [4]H. Angm, “Common vector obtained from linearly independent speech vectors by using LPC parameters,” graduation project, Elect. Electron. Eng. Dept., Osmangazi Univ., Eskisehir, Turkey, 1995. [5]Hakan Cevikalp and Mitch Wilkes,”Discriminative Common Vectors for Face Recognition”pattern analysis and machine intelligence,vol.27,NO.1, 2005 [6]L. Rabiner and B. H. Juang,”Fundamentals of Speech Recognition”, Englewood Cliffs,NJ:prentice-Hall,1993. [7]Frederick Jelinek ,”Statistical methods for speech recognition ”, Mass.:/MIT Press,Cambridge,1998 [8]李宗寶,吳宗憲。” 探討Kmean之共同向量法應用於國語數字辨識”。碩士論文,國立中興大學應用數學研究所,台中,2005。 [9]王小川。”語音訊號處理”。台北市:全華,2004。 [10]李宗寶,林靖剛。”利用Multiple Common Vector 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,2006。
摘要: 
摘要
本篇論文主要是利用「Discriminative Common Vectors法」,來探討國語數字0 ~ 9的語音辨識,「Discriminative Common Vectors法」原先是使用在臉部辨識的方法之一,在辨識率上有一定的水準,所以本人試著套用此方法在語音辨識上,利用臉部辨識的方法來做語音的辨識,找出最相似的語音,看效果是否會較好。
本實驗中的語音資料庫,由21人(12名男生,9名女生)所建構,每人唸0 ~ 9各三次,前兩次做為訓練語音模型用,第三次則做為特定語者辨識時測試用,接著對所有語音資料做前處理,並建構模型來做為辨識比對。
本論文所要討論影響辨識率的因子有音框取樣個數、子母音的比例分配兩種。在特定語者的辨識結果最高為取樣音框個數為200,子母音分配比例為1.5:8.5的狀況下,辨識率最高為92.86%。
URI: http://hdl.handle.net/11455/17884
其他識別: U0005-2606200712432100
Appears in Collections:應用數學系所

Show full item record
 

Google ScholarTM

Check


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