Identity Verification with Speech Recognition : A Study

Abstract

Since verification and authentication based on speech recognition are important develop-ments in the field of information security, this study revolves around the topic of speech-based Identity verification. The goal was to create a speech recognition system that can be embedded to an existing user authentication system.

For this purpose, the study relied on a free speech modelling tool and modelled a limited dictionary for speech recognition. Speech samples were collected for training and testing purposes. The training sample was collected from three respondents and the testing data from seven, including those from previous three respondents. A freely available tool, HTK Toolkit, was used to train the model with speech samples recorded from respondents. The results showed that the accuracy of the model was dependent on whether the training da-taset included the users’ speech or not.

The study supports the significance of real use of the model. However, the scope of the study is not large enough for authentication and verification system.