摘要
在大词汇量连续语音识别应用中,优质的语音训练语料是所有识别工作的基础和前提,能否挑选出覆盖更多语音现象的语料是提高语音识别性能的关键。该文在多种维吾尔文口语化传播平台中采集了大量口语句子语料,并考虑协同发音的影响和常用词的适用性,根据评估函数对语料筛选。经过筛选后的语料包含的三音子更加均衡和高效,囊括的语音现象更加全面,为训练准确而牢靠的语音模型打下了稳固的根基。
A good speech training corpus is essential for the wide application of continuous speech recognition. Therefore, whether more multiple yoice phenomena are covered in the corpus is of substantial importance to improve the performance of speech recognition. In this paper, we collect a large number of spoken corpus sentences from a variety of Uigl^ur spoken language communication platforms. Then, we refine the corpus according to the evaluation function considering the effect of co-articulation and applicability of the common words. The final corpus contain mot more balanced and efficient tri-phones, covering more phonetic phenomena, which lays a solid foundation for training a much accurate and reliable acoustic model.
出处
《中文信息学报》
CSCD
北大核心
2015年第2期118-124,共7页
Journal of Chinese Information Processing
基金
国家973重点基础研究计划项目(2014CB340506)
国家自然科学基金(61363063)
新疆维吾尔自治区科技计划项目(201312104)
关键词
维吾尔语
语音识别
语料库
三音子
Uighur Language
speech recognition
database
Triphone
