| 一种语音端点检测算法在DSP上的实现 |
| |
| 引用本文： | 胡钢.一种语音端点检测算法在DSP上的实现[J].鞍山师范学院学报,2009,11(6):44-47. |
| |
| 作者姓名： | 胡钢 |
| |
| 作者单位： | 鞍山师范学院,财经系,辽宁,鞍山,114007 |
| |
| 摘 要： | 介绍了一种对背景噪音进行动态估计的汉语语音端点检测算法，较一些常规的端点检测方法如基于能量的端点检测方法，具有更高的精度和鲁棒性．并用TI的TMS320VC5410 DSP来实现一个带端点检测的语音识别系统．通过进行汉语孤立词语音识别实验，表明这种算法是有效的，能得到较高的识别率．
|
| 关 键 词： | 语音识别 端点检测 DSP HMM |
DSP Implementation of Speech Recognition Based on Endpoint Detection of Speech |
| |
| HU Gang.DSP Implementation of Speech Recognition Based on Endpoint Detection of Speech[J].Journal of Anshan Teachers College,2009,11(6):44-47. |
| |
| Authors: | HU Gang |
| |
| Institution: | HU Gang ( Anshan Normal University ,Anshan Liaoning 114007 ,China) |
| |
| Abstract: | The author proposes one algorithm that detects the beginning and ending boundaries of speech based on dynamic estimating environmental noises. This algorithm can work well with more accurate and robust than time-domain-energy-based endpoint detection. A speaker-independent speech recognition system that contains endpoint detection of speech is developed based on TMS320VC5410 Evaluation Module （EVM）. The experiment based on isolated-word speech-recognition shows that：this algorithm contributes to the increase of recognition rate. |
| |
| Keywords: | DSP HMM |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
