摘要
论述了强背景噪声环境下利用模糊分类算法对大词汇量的汉语语音进行分类识别，根据噪声中的汉语语音特点，采用有边界的交叉分类和无边界的模糊分类相结合的措施，较成功地解决了强噪声环境下的汉语语音分类．该算法在７５ｄＢ的强噪声环境下，对５００个汉语短语进行识别，其识别精度达到９５％，取得了较好的识别效果．
A method utilizing fuzzy classified algorithms for the speech recognition of the Chinese language with a large vocabulary in a strong noise environment was proposed. Accoding to the phonetic characteristics of the Chinese language, the boundary crossing classified and the boundaryless fuzzy classified algorithems were successfully used in classifying the Chinese language with a large vocabulary in a strong niose environment. Good results were achieved in applying these algorithms to speech recognition system.The correctly recognizing rate is 95% of 500 Chinese phrases under test in the 75dB noise enviroment.
出处
《北京理工大学学报》
EI
CAS
CSCD
1997年第6期686-690,共5页
Transactions of Beijing Institute of Technology
基金
国家"八五"预研项目
关键词
强噪声环境
汉语语音识别
模糊分类法
语音识别
strong noise environment
Chinese speech recognition
fuzzy classifying methods
boundary crossing classifying method
bounderyless fuzzy classifying method
