维吾尔语连续语音识别技术研究

Research on Uyghur Continuous Speech Recognition Technology

维吾尔语连续语音识别技术研究主要阐述维吾尔语连续语音的识别技术,主要包括声学模型和语言模型。在声学模型中,主要介绍基于隐马尔可夫模型(Hidden Markov Model,HMM)的维吾尔语连续语音识别声学建模。在语言模型中,主要对比基于文法和基于统计这两种方法的优劣。

Mainly elaborates the technology of the continuous speech recognition each aspect, the main is the acoustic model and the language model technology. In the acoustic model, mainly introduces the acoustic modeling of Uyghur continuous speeches recognition, which based on Hidden Markov Model. But in the language model, it primarily contrast the quality of these two method, which are based on grammar and statistics.