昆明理工大学学报（自然科学版）
昆明理工大學學報（自然科學版）
곤명리공대학학보（자연과학판）
JOURNAL OF KUNMING UNIVERSITY OF SCIENCE AND TECHNOLOGY（SCIENCE AND TECHNOLOGY）
2014年
5期
47-53
,共7页
余炜%周娅%万代立%杨喜敬
餘煒%週婭%萬代立%楊喜敬
여위%주아%만대립%양희경
语音识别%孤立词识别%彝语%Mel频率倒谱系数%DTW
語音識彆%孤立詞識彆%彝語%Mel頻率倒譜繫數%DTW
어음식별%고립사식별%이어%Mel빈솔도보계수%DTW
speech recognition%isolated-word recognition%Yi language%MFCC%DTW
彝语的语音识别与处理是语音信号处理领域的一个新方向。本文在深入分析彝语特点的基础上，对用于汉语孤立词识别的经典端点检测算法进行了改进。探讨了动态时间规整算法（Dynamic Time Warping，DTW）和高效动态时间规整算法（Efficient DTW，EDTW）在彝语孤立词识别中的应用，并提出了两种新算法：基于音节个数的动态时间规整算法（Syllable Number Based Improved DTW，SDTW）和基于音节个数的高效动态时间规整算法（Syllable Number Based Im-proved EDTW，SEDTW），对特定的彝语语音信号进行识别。实验结果表明，所提出的两种新算法分别在信号识别率和识别时间上比经典算法优越。
彝語的語音識彆與處理是語音信號處理領域的一箇新方嚮。本文在深入分析彝語特點的基礎上，對用于漢語孤立詞識彆的經典耑點檢測算法進行瞭改進。探討瞭動態時間規整算法（Dynamic Time Warping，DTW）和高效動態時間規整算法（Efficient DTW，EDTW）在彝語孤立詞識彆中的應用，併提齣瞭兩種新算法：基于音節箇數的動態時間規整算法（Syllable Number Based Improved DTW，SDTW）和基于音節箇數的高效動態時間規整算法（Syllable Number Based Im-proved EDTW，SEDTW），對特定的彝語語音信號進行識彆。實驗結果錶明，所提齣的兩種新算法分彆在信號識彆率和識彆時間上比經典算法優越。
이어적어음식별여처리시어음신호처리영역적일개신방향。본문재심입분석이어특점적기출상，대용우한어고립사식별적경전단점검측산법진행료개진。탐토료동태시간규정산법（Dynamic Time Warping，DTW）화고효동태시간규정산법（Efficient DTW，EDTW）재이어고립사식별중적응용，병제출료량충신산법：기우음절개수적동태시간규정산법（Syllable Number Based Improved DTW，SDTW）화기우음절개수적고효동태시간규정산법（Syllable Number Based Im-proved EDTW，SEDTW），대특정적이어어음신호진행식별。실험결과표명，소제출적량충신산법분별재신호식별솔화식별시간상비경전산법우월。
The recognition and processing of Yi language is a new direction of speech signal processing field. Based on deep analysis of the characteristics of Yi,the classic endpoint detection algorithm for Chinese isolated-word are improved.The applications of dynamic time warping (DTW)algorithm and efficient dynamic time warping(EDTW)algorithm for Yi isolated -word are discussed,and two novel algorithms for this purpose are proposed,which are Syllable Number Based Improved DTW and Syllable Number Based Improved EDTW.Ex-perimental results show that the proposed algorithms have better performance on the recognition rate and time compared with classic methods.
