中文说明:实现简单的声音识别。本程序为课程实验验证开发,所实现的声纹识别算法难免有错误及不妥之处! 提醒:因为本程序是简略的验证程序,建议不要在正式项目中直接使用本程序中的声纹识别代码。VoicePrint的特点是: - 语音特征向量:采用 MFCC系数(梅尔倒谱系数)- 模式匹配模型:采用 GMM 模型(高斯混合模型)
English Description:
Realize simple voice recognition. This program is developed for the experimental verification of the course, the voiceprint recognition algorithm is inevitable to have errors and defects! &Reminder: because this program is a simple verification program, it is not recommended to use the voiceprint identification code directly in the formal project. The features of voiceprint are: & nbsp; - speech feature vector: MFCC coefficient (Mel cepstrum coefficient) - pattern matching model: GMM model (Gaussian mixture model)