Research on Text-independence for Voiceprint Recognition System
Voiceprint recognition was divided into text-dependence and text-independence according to the recognition pattern. In view of voiceprint recognition technique, the problem on the text-independent voiceprint recognition was resear ched. Multi-feature vectors combination recognition algorithm was proposed and implemented to improve the system recognition rate. This algorithm took the combination of LPCC and MFCC as the feature parameters which were also put into use in voiceprint recognition system. The result of experiments expatiates that the parameters featured combination has much better recognition effect than a single kind of parameters used, because relative voice information with features and human hearing characteristics are fully used.
NETL
NSTL
万方数据
维普
