A Hybrid Mel Frequency Cepstral Coefficients and Bayesian Gaussian Mixure Model for Voice based Authentication Websites

Consumers of today can interact with technology to complete a variety of simple tasks using voice recognition, a new technology. Because of advancements in technology, voice recognition systems help identify people by their speech. It provides quick and extremely secure identification and authentication with a minimum of speech. Create a voice-based website authentication system that is user-friendly as part of this research to access websites. As part of the registration process, users' speech signals are initially gathered for this study. It is recommended that the Hybrid Mel Frequency Cepstral Coefficients and Bayesian Gaussian Mixture Model be used to train the speech authentication system (HMFCC-BGMM). In this instance, features from the voice are extracted using MFCC, and the system is trained using BGMM. According to the experimental findings, the proposed HMFCC-BGMM has a high accuracy of 94.5% and a lower error rate than previous models.