QQ客服 官方微博 反馈留言

	适于嵌入式英语发音评价系统的英美音融合方法
	梁维谦 ; 赵坤 ; 刘润生 ; LIANG Weiqian ; ZHAO Kun ; LIU Runsheng
	2010-07-15 ; 2010-07-15
会议名称	第十届全国人机语音通讯学术会议暨国际语音语言处理研讨会论文摘要集 ; Abstract Book of the 10~(th) National Conference on Man-Machine Speech Communication (NCMMSC'2009) and International Symposium on Speech and Language Processing ; 第十届全国人机语音通讯学术会议暨国际语音语言处理研讨会 ; The 10~(th) the National Conference on Man-Machine Speech Communication (NCMMSC'2009) and International Symposium on Speech and Language Processing ; 中国甘肃兰州 ; CNKI ; 中国中文信息学会语音信息专业委员会、中国声学学会语言、听觉和音乐声学分会、中国语言学会语音学分会
关键词	计算机辅助语言学习 嵌入式应用 发音质量评价 模型融合 computer assisted language learning embedded applications pronunciation assessment model merging H311.
其他题名	Model Merging of British and American Accents for Embedded Pronunciation Scoring Applications
中文摘要	英音和美音是最为常见的英语目标口音。本文针对英语学习者存在的英美口音混杂现象,提出一种英美口音模型融合方法,提高发音质量评价性能,并压缩声学模型规模以适于嵌入式实现。该方法基于替代概率将声学模型划分为可替代模型、可融合模型和孤立模型。抛弃可替代模型,保留孤立模型,基于模型插值与模型裁剪归并可融合模型。发音质量评价实验结果表明:融合模型与单口音模型相比,说话人级别的相关性提高了14.1%;融合模型与混合模型的性能相近,高斯分量数目压缩了10.7%。; American English and British English are both very popular. In this paper, we propose a new model merging method which can cover both American and British accents to improve the performance of pronunciation scoring with smaller model size. Acoustic models are classified into tree class-replaceable models, merging models, isolated models, based on the rank of the substituting probability. We discard the replaceable models, keep the isolated models, and merge the merging models by using model interpolation and pruning. Experimental results show, that by using this merged models the speaker level correlation between machine scores and human scores improves about 14.1% compared to the single-accent models, and the number of Gaussian mixtures reduces 10.7% compared to the combined models.; 国家高技术研究发展计划(863计划)重点项目(2008AA010700)
会议录出版者	清华大学
语种	中文 ; 中文
内容类型	会议论文
源URL	[http://hdl.handle.net/123456789/69777]
专题	清华大学
推荐引用方式 GB/T 7714	梁维谦,赵坤,刘润生,等. 适于嵌入式英语发音评价系统的英美音融合方法[C]. 见:第十届全国人机语音通讯学术会议暨国际语音语言处理研讨会论文摘要集, Abstract Book of the 10~(th) National Conference on Man-Machine Speech Communication (NCMMSC'2009) and International Symposium on Speech and Language Processing, 第十届全国人机语音通讯学术会议暨国际语音语言处理研讨会, The 10~(th) the National Conference on Man-Machine Speech Communication (NCMMSC'2009) and International Symposium on Speech and Language Processing, 中国甘肃兰州, CNKI, 中国中文信息学会语音信息专业委员会、中国声学学会语言、听觉和音乐声学分会、中国语言学会语音学分会.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。