On the application of speaker verification to electronic-commerce

As the electronic-commerce (EC) market grows rapidly, user authentication using biometric technology draws more and more attention. We investigated the possibility of speaker verification as a biometric authentication system in EC. We identified some research issues in enrollment sessions and test sessions of the speaker verification system and then compared the performance of several representative speaker verification algorithms. The experimental result shows that the system with a hidden Markov model (HMM) has the best performance with the lowest equal error rate (EER) of 0.87%.