首页 | 本学科首页   官方微博 | 高级检索  
     检索      

语音技术在少数民族语言的应用研究
引用本文:达瓦.伊德木草,木合亚提.尼亚孜别克,吾守尔.斯拉木.语音技术在少数民族语言的应用研究[J].新疆大学学报(理工版),2014(1):88-96.
作者姓名:达瓦.伊德木草  木合亚提.尼亚孜别克  吾守尔.斯拉木
作者单位:[1]新疆维吾尔自治区多语种技术重点实验室,新疆乌鲁木齐830046 [2]新疆大学信息科学与工程学院,新疆乌鲁木齐830046
基金项目:国家自然科学基金(61163030)及自治区科技支疆项目(201291116)资助.
摘    要:本文研讨缺乏语言资源的民族语言(如维吾尔语)中如何引用语音技术、开发应用系统问题.提出基于GMM-UBM混合SVM技术方法实现实用性说话人识别系统,通过小语料人工标注语音语料预选高精度声学根(seed)模型、再引导大语料训练生成鲁棒性声模提高连续语音识别精度实现汉民会话语音翻译系统.对维吾尔语70人发话电话语音识别实验结果显示,基于GMM-UBM -SVM方法的不特定说话人识别实验其正确识别率为94.3%,比先行GMM-UBM方法精度提升3%;基于seed声模HTK-Julius技术的维吾尔语连续语音识别实验,其识别率为72.5%,比直接使用语音文本对齐语料单靠HTK实现识别方法(63.2%)精度提高9.3%;同时本研究讨论基于Moses技术的汉维医院门诊会话语音翻译系统预测Blue值达到了57.7%.

关 键 词:语音技术  民语  说话人识别  连续语音识别  语音翻译

An Applied Research of Speech Technology in Resource-deficient Languages
Institution:Dawa Yidemucao, Muheyati Niyazibeke, Wushour Silamu(1. Xinjiang Key Laboratory of Multi-Language Information Technology, Urumqi Xinjiang 830046, China; 2. College of Information Science and Engineering, Xinjiang University, Urumqi Xinjiang 830046, China)
Abstract:In this paper, we report our recent researches in speech techniques, such as speaker identification (SI), continuous speech recognition (CSR) and speech translation (ST), among the minority languages spoken in China, like Uyghur, Khasak and Mongolian. We tried an approach GMM-UBM-SVM for real calling SI, and HTK-Julius for CSR by Uigur and Mongolian, and Moses software for ST for medical service. Experimental results show that accuracies of 94.3%for SI with GMM-UBM-SVM, and of 72.5%for CSR by Uygur, and showed a Blue value of 57.7%for the medical service ST by Uyghur using Moses software.
Keywords:GMM-UBM-SVM  Speech technology  speaker identification  continuous speech recognition  speech translation  GMM-UBM-SVM
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号