| 基于HTK的普米语孤立词的语音识别 |
| |
| 作者单位： | ;1.云南民族大学数学与计算机科学学院;2.云南民族大学云南省高校物联网应用技术重点实验室 |
| |
| 摘 要： | 基于HTK的普米语孤立词的识别,通过对特定发音人和非特定发音人的语音进行特征分析、预处理、提取MFCC特征参数,从而建立每个普米词的HMM模型,最后采用Viterbi算法进行模型的训练和匹配.实验表明:对特定发音人和非特定发音人,随着普米词个数的增加,识别率虽有所下降,但是变化很小,识别率均超过了95.00%,识别效果很好.语料的信噪比和发音人的发音稳定性是导致识别率下降的重要原因.
|
| 关 键 词： | HTK Viterbi算法 普米语 HMM 语音识别 |
Speech recognition of isolated word speech of primi language based on HTK |
| |
| Institution: | ,School of Mathmatics and Computer Science,Yunnan Minzu University,Key Laboratory of IOT Application Technology of Universities in Yunnan Province,Yunnan Minzu University |
| |
| Abstract: | This paper talks about the recognition of isolated words in Primi language based on HTK. We get the feature analysis,pretreatment,extraction of MFCC feature parameters through the voice of the specific pronunciation and the different pronunciation. So it can establish the HMM model of each of the words of Primi. Viterbi algorithm is used to train the model and match to achieve speech recognition of Primi language. Experimental results show:The recognition rate has decreased along with the increase of the number of Primi words to the specific pronunciation and the different pronunciation. But the change is very small,the recognition rate is more than 95% and the recognition effect is very good. The degree of stability of the pronunciation and signal noise ratio are two important factors to lead the decline of recognition rate. |
| |
| Keywords: | HTK Viterbi algorithm Primi language HMM speech recognition |
| 本文献已被 CNKI 等数据库收录！ |
|
