A dynamic-threshold approach to text-dependent speaker recognition using principles of immune system

This paper focuses on implementation of a speaker recognition system, where a user will require to provide a particular word in order to be verified as a particular user. Further, the principles of immune system are used to model the problem mentioned above. The topics covered in this paper describe the methodology of a text-dependent speaker verification process modeled using an immune-engineering framework. The implementation follows a free software approach by using Python programming language rather than conventional Matlab, which is proprietary by license. The developed system is tested in a personal computer using web cam microphone and in a noise-less environment.