论文部分内容阅读
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。
Uyghur continuous speech recognition technology research mainly focuses on Uyghur continuous speech recognition technology, including acoustic model and language model. In the acoustic model, Uyghur continuous speech recognition acoustic modeling based on Hidden Markov Model (HMM) is mainly introduced. In the language model, the main comparison is based on grammar and statistics based on the merits of these two methods.