基于声纹识别的身份认证算法研究

With the development of technology, the authentication method based on voiceprint recognition has become a research hotspot. Current voiceprint recognition technology uses a single feature or a combination of stitching features to build a recognition model, which is less robust to complex environments. In order to address this problem, this paper first proposes a Var-MFCC feature based on MFCC features. Then, a hierarchical multi-feature fusion set is constructed using voiceprint features of different categories, and based on it，a multi-feature weighted fusion algorithm based on Fisher ratio is proposed, Finally, based on the above features and fusion mechanism, this paper constructs an identity authentication model based on voiceprint recognition.. The experimental result of the speech data set in a non-ideal environment shows that the multi-feature fusion identity authentication model constructed in this paper has better performance and processing speed than the baseline mode.