|
|||||
|
|
| 汉语语音识别中的区分性声调建模方法 | |
| 引用本文： | 黄浩,朱杰,哈力旦.汉语语音识别中的区分性声调建模方法[J].计算机工程与应用,2009,45(11):178-182. |
| 作者姓名： | 黄浩 朱杰 哈力旦 |
| 作者单位： | 1. 上海交通大学,电子工程系,上海,200240;新疆大学,信息科学与工程学院,乌鲁木齐,830046 2. 上海交通大学,电子工程系,上海,200240 3. 新疆大学,电气工程学院,乌鲁木齐,830046 |
| 摘 要： | 提出从特征提取参数、模型参数对隐马尔可夫声调模型进行区分型训练,来提高声调识别率;提出模型相关的权重对谱特征模型和声调模型的概率进行加权,并根据最小音子错误区分性目标函数对权重进行训练,来提高声调模型加入连续语音识别时的性能。声调识别实验表明区分性的声调模型训练以及特征提取方法显著提高了声调识别率。区分性模型权重训练能够在声调模型加入之后进一步连续语音识别系统的识别率。 |
| 关 键 词： | 区分性训练 声调建模 汉语语音识别 特征提取 |
| 收稿时间： | 2008-9-22 |
| 修稿时间： | 2008-11-26 |
Tone modeling based on discriminative training for Mandarin speech recognition |
|
| HUANG Hao,ZHU Jie,HA Li-dan.Tone modeling based on discriminative training for Mandarin speech recognition[J].Computer Engineering and Applications,2009,45(11):178-182. | |
| Authors: | HUANG Hao ZHU Jie HA Li-dan |
| Affiliation: | HUANG Hao1,2,ZHU Jie1,HA Li-dan31.Department of Electronic Engineering,Shanghai Jiao Tong University,Shanghai 200240,China 2.Department of Information Science , Engineering,Xinjiang University,Urumqi 830046,China 3.Department of Electrical Engineering,China |
| Abstract: | To improve tone recognition accuracy,discriminative training in both feature and model parameters for hidden Markov model based tone modeling is proposed.When incorporating tone models into continuous speech recognition,discriminative model weight training is presented.Acoustic and tone model distributions are scaled by model dependent weights trained by the mini-mum phone error criterion.Experiments show tone recognition and large vocabulary continuous speech recognition accuracy can be considerably improv... |
| Keywords: | discriminative training tone modeling Mandarin speech recognition feature extraction |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |
| 点击此处可从《计算机工程与应用》浏览原始摘要信息 | |
| 点击此处可从《计算机工程与应用》下载全文 | |
