| Linear Discriminant Analysis and Kernel Vector Quantization for Mandarin Digits Recognition |
| |
| 作者姓名： | 赵军辉 谢湘 匡镜明 |
| |
| 作者单位： | SchoolofInformationScienceandTechnology,BeijingInstituteofTechnology,Beijing100081,China |
| |
| 基金项目： | SponsoredbyR&DCooperationProjectionBetweenEricssonandBeijingInstituteofTechnology |
| |
| 摘 要： | Linear discriminant analysis and kernel vector quantization are integrated into vector quantization based speech recognition system for improving the recognition accuracy of Mandarin digits. These techniques increase the class separability and optimize the clustering procedure. Speaker-dependent (SD) and speaker-independent (SI) experiments are performed to evaluate the performance of the proposed method. The experiment results show that the proposed method is capable of reaching the word error rate of 3.76 % in SD case and 6.60 % in SI case. Such a system can be suitable for being embedded in personal digital assistant(PDA), mobile phone and so on to perform voice controlling such as digit dialing, calculating, etc.
|
| 关 键 词： | 语音识别 阿拉伯数字 直线判别式 无线电引导 量子化 |
| 收稿时间： | 2003/8/18 0:00:00 |
Linear Discriminant Analysis and Kernel Vector Quantization for Mandarin Digits Recognition |
| |
| ZHAO Jun-hui,XIE Xiang and KUANG Jing-ming.Linear Discriminant Analysis and Kernel Vector Quantization for Mandarin Digits Recognition[J].Journal of Beijing Institute of Technology,2004,13(4):385-388. |
| |
| Authors: | ZHAO Jun-hui XIE Xiang and KUANG Jing-ming |
| |
| Affiliation: | School of Information Science and Technology, Beijing Institute of Technology, Beijing 100081, China |
| |
| Abstract: | Linear discriminant analysis and kernel vector quantization are integrated into vector quantization based speech recognition system for improving the recognition accuracy of Mandarin digits. These techniques increase the class separability and optimize the clustering procedure. Speaker-dependent (SD) and speaker-independent (SI) experiments are performed to evaluate the performance of the proposed method. The experiment results show that the proposed method is capable of reaching the word error rate of 3.76% in SD case and 6.60 % in SI case. Such a system can be suitable for being embedded in personal digital assistant(PDA), mobile phone and so on to perform voice controlling such as digit dialing, calculating, etc. |
| |
| Keywords: | linear discriminant analysis kernel vector quantization speech recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《北京理工大学学报(英文版)》浏览原始摘要信息 |
|
点击此处可从《北京理工大学学报(英文版)》下载全文 |
