摘要
采用平均幅度差法、自相关函数法和简单逆滤波器跟踪法相结合的方法计算语音信号的基音频率.根据声调模式的不同,采用基于规则的方法对声调进行识别,对连续数字串识别中一些易混淆的数字对进行区分,从而达到提高数字串识别系统性能的目的.
According to the extracted pitch curve, it can discriminate between four Mandarin tones. A composite algorithm for pitch extraction was promoted which integrates AMDF, auto correlation and simple inverse filtering trucking (SIFT) algorithm by using some rules. In the Mandarin continuous digit string recognition system, the tones are used to discriminate some confusing digit pairs, which can improve the system's recognition rate.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2004年第2期185-188,共4页
Journal of Shanghai Jiaotong University
基金
上海市科学技术委员会基础研究基金(01JC14033)
美国贝尔实验室上海分部资助项目
关键词
语音识别
声调
数字识别
Computer simulation
Correlation methods