基于MFCC参数和VQ的说话人识别系统  

Speaker recognition system using MFCC features and vector quantization

作　　者：王伟[1] 邓辉文[1] 

机构地区：[1]西南大学计算机与信息科学学院,重庆400715

出　　处：《仪器仪表学报》2006年第z3期2253-2255,共3页Chinese Journal of Scientific Instrument

摘　　要：采用能够反映人对语音的感知特性的Mel频率倒谱系数(MFCC)作为特征参数,以及为避免时间规整问题采用矢量量化技术开发的说话人识别系统。MFCC主要的是模拟人耳的听觉过程,相对于其它参数它对语音波形的变化不敏感,更加稳定,系统取得很好的识别结果,实验表明系统训练和识别的计算量和存储量都比较低。This paper presents a speaker recognition system which uses Mel-frequency cepstrum coefficients (MFCC) reflected person's apperceptionnature as feature parameters. To avoid the problem of DTW, we adopt the vector quantization technology . The main purpose of the MFCC processor is to mimic the behavior of the human ears. Rather than the speech waveforms themselves, MFCC is shown to be less susceptible ,more stable to other variations. The result of recognition is goodand the requirement of computation and storage is quitelow.

关 键 词：说话人识别 MEL倒谱系数 矢量量化 

分 类 号：TH7-55[机械工程—仪器科学与技术]