基于改进LPCC和MFCC的汉语耳语音识别  

Identification of Chinese whispered speech based on modified LPCC and MFCC

作　　者：荣薇[1] 陶智[1] 顾济华[1] 赵鹤鸣[2] 

机构地区：[1]苏州大学物理科学与技术学院,江苏苏州215006 [2]苏州大学电子信息学院,江苏苏州215006

出　　处：《计算机工程与应用》2007年第30期213-216,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(the National Natural Science Foundation of China under Grant No.60572076);江苏省高校自然科学研究计划项目(No.05JKB510113)。

摘　　要：以提高汉语耳语的识别率为目的,提出了将MFCC、LPCC及它们各自的动态参数等多种特征有效结合进行耳语识别的方法。实验结果说明了LPCC、MFCC结合动态参数可作为汉语耳语音识别的特征参数,且它们的结合提高了系统的识别率,在小字库内得出的识别率为94.5%。In order to improve the rate of identification of the whispered speech,this paper puts forward a method that combining efficiently more features such as MFCC and LPCC and their dynamic parameters and so on to acquire better recognition performance.The experimental results demonstrate that the LPCC and MFCC combine with the dynamic parameters features can be used as efficient parameters for the Chinese whispered speech recognition,and their combination improve the performance of the whispered speaker recognition system.A recognition rate of 94.5% is obtained when a small Chinese isolated word database is tested.

关 键 词：耳语音 语音识别 LPCC参数 MFCC参数 

分 类 号：TN912[电子电信—通信与信息系统]