跳到主要內容

臺灣博碩士論文加值系統

(3.236.84.188) 您好!臺灣時間:2021/08/02 20:15
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:李柏蒼
論文名稱:自發性國語語音辨識
論文名稱(外文):Spontaneous Mandarin Speech Recognition
指導教授:王逸如
學位類別:碩士
校院名稱:國立交通大學
系所名稱:電信工程系所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
畢業學年度:96
語文別:中文
論文頁數:48
中文關鍵詞:自發性語音語音辨識
外文關鍵詞:Spontaneous SpeechSpeech Recognition
相關次數:
  • 被引用被引用:5
  • 點閱點閱:151
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
自發性語音是最接近人們在自然情況下的語音,較能實際應用於日常生活中,因此漸趨於重要。本論文將先用國語sub-syllable HMM 建立聲學模型,從錯誤分析知道,Uncertain 與Particle 常與411 syllable 互相辨識,將Uncertain 不加入辨識器,可提高辨識率,Particle 在聲學上,十分相似411 syllable,只差異語意,聲學模型是難以區別。再以Syllable HMM 建立聲學模型,將可提高辨識率
3.3%,使用skip state 改善Deletion 錯誤,改善並不如預期,從錯誤分析知道Deletion 錯誤大部分由Syllable Contraction 造成。
The spontaneous speech is most close to the speech in natural cases of people and can relatively apply to daily life actually, so become more and more important gradually. The first sets up acoustics model with mandarin sub-syllable HMM. From the error analysis, the recognizing device doesn’t often distinguish between Uncertain and 411 syllable, and between Particle and 411 syllable. Do not put Uncertain model into the recognizing device, can improve the recognizing rate. Particle is in acoustics, very similar 411 syllable, and the language purpose of Particle is only different from the language purpose of 411 syllable and this is difficult to distinguish between their acoustics models. The second sets up acoustics model with mandarin sub-syllable HMM, and then can improve 3.3% of recognizing rate, and uses skip state to improve Deletion error, and then that does not improve so well as expectancy. From Deletion error analysis, we know Syllable Contraction causes the most of Deletion error.
第一章 緒論................................................1
1.1 研究動機...............................................1
1.2 研究方向...............................................1
1.3 章節概要...............................................2
第二章 現代漢語口語對話語料庫之介紹與統計..................3
2.1 MCDC 之簡介............................................3
2.2 音檔格式說明...........................................4
2.2.1 原始音檔處理方式.....................................4
2.2.2 音檔後處理方式.......................................4
2.2.3 音檔格式比較.........................................5
2.3 文字轉寫格式說明.......................................5
2.4 MCDC 口語之轉寫標示....................................7
2.4.1 非語音現象(Paralinguistic Phenomena) ................7
2.4.2 不確定字/音(Uncertain)...............................7
2.4.3 語助詞(Marker) ......................................8
2.4.4 感嘆詞 (Particle)....................................8
2.4.5 語言轉換 (Code Switching) ...........................8
2.5 MCDC 之相關統計........................................8
第三章 MCDC 基本語音辨識實驗..............................11
3.1 訓練語料與測試語料....................................11
3.2 聲學模型的建立........................................12
3.2.1 特徵參數............................................12
3.2.2 國語sub-syllable HMM 之建立.........................12
3.3 實驗結果..............................................14
3.3.1 錯誤分析............................................14
3.3.1.1 Uncertain 取代型錯誤分析..........................15
3.3.1.2 Particle 取代型錯誤...............................16
3.3.1.3 Particle model 與同音411 syllable model 之分析....17
3.3.1.4 Particle 與同音411 syllable 的duration 分佈.......21
3.3.1.5 Particle 前後silence 的分析.......................22
3.3.1.6 加上前後silence 後,Particle 與同音411 syllable 的duration分佈..............................................24
3.3.1.7 調整Particle 與411 syllable 發生的機率............26
3.4 檢查語料錯誤..........................................29
第四章 Syllable HMM 之建立................................31
4.1 Syllable HMM .........................................31
4.2 Skip State Syllable HMM ..............................33
4.2.1 K-L Distance........................................35
4.3 Deletion 錯誤分析.....................................37
第五章 結論與未來展望.....................................43
5.1 結論..................................................43
5.2 未來展望..............................................43
參考文獻..................................................45
附錄..................................................47
【1】B.H. Juang and S. Furui, Automatic Recognition and Understanding of Spoken Language - A First Step Toward Natural Human–Machine Communication,Proc, IEEE, 88, 8, pages 1142-1165, 2000.
【2】Rabiner, L.R. and Juang, B.H., Fundamentals of speech Recognition, New Jersey, Prentice-Hall,Inc.,1993.
【3】Lin, C.-K., et al., “Important and New Features with Analysis for Disfluency Interruption Point (IP) Detection in Spontaneous Mandarin Speech”, in Proc. of DiSS, 2005.
【4】Lin, C.-K. & Lee, L.-S. Improved Spontaneous Mandarin Speech Recognition by Disfluency Interruption Point (IP) Detection Using Prosodic Features. Proc. Eurospeech’05.
【5】吳維彥,「應用不定長度特徵之條件隨機域於口語不流暢語流修正模型」,國立成功大學資訊工程學系碩士論文,民國九十五年六月。
【6】羅應順,「自發性中文語音基本辨認系統之建立」,國立交通大學電信工程學系碩士論文,民國九十四年六月。
【7】徐文翰,「自發性對話語音辨認之初步研究」,國立交通大學電信工程學系碩士論文,民國九十三年七月。
【8】曾淑娟,劉怡芬,現代漢語口語對話語料庫標註系統說明,中央研究院語言學研究所籌備處,民國九十一年九月。
【9】S. Young, G. Evermann, T. Hain, D. Kershaw, G. Moore, J. Odell, D. Ollan, D. Povey, V. Valtchev, P. Wooland, 「The HTK Book(for HTK version 3.4)」.
【10】S. Kullback and R. Leibler, On information and sufficiency, Ann. Math. Statist. , vol. 22, pp. 79–86, 1951. 46
【11】Shu-Chuan Tseng, Contracted Syllables in Mandarin: Evidence from Spontaneous Conversations, Academia Sinica
【12】Shu-Chuan Tseng, Syllable Contractions in a Mandarin Conversational Dialogue Corpus, Institute of Linguistics, Academia Sinica, Taiwan
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top