摘要
近年来大词汇量连续语音识别技术得到了迅速的发展,国内外研究机构加大了对汉语和英语语音识别技术的研究,然而,维吾尔语语音识别技术的研究工作最近才起步。建立了面向大词汇量的维吾尔语语音语料库,研究了维吾尔语声学模型和语言模型建模技术、解码技术,进行了面向大词汇量的维吾尔语连续语音识别实验。对维吾尔语大词汇量连续语音识别技术进一步发展中存在的问题进行了讨论。
The technology of Large Vocabulary Continuous Speech Recognition(LVCSR) has developed quickly, and many scientific institutions have reinforced the speech recognition research on the Mandarin Chinese and English. However, the study of Uyghur speech recognition technology has started recently. This paper introduces the research on main aspect of Uyghur LVCSR system, such as construction of Uyghur speech corpus, acoustic and language modeling techniques, decoding techniques, and performed experiments for Uyghur LVCSR. At the end, the issues affecting Uyghur LVCSR system are discussed in detail.
出处
《计算机工程与应用》
CSCD
2013年第9期115-119,共5页
Computer Engineering and Applications
基金
国家自然科学基金(No.61063024)
新疆大学联合科研项目(No.XY110122)
关键词
维吾尔语
语音语料库
大词汇
识别技术
Uyghur language
speech corpus
large vocabulary
recognition technology