VOICE RECOGNITION FOR SECURITY PURPOSES

This report presents the implementation of Voice Recognition System for Security
Purposes. The scope of study is to understand the existing algorithms and develop a
brand new algorithm which is better in terms of its reliability and accuracy. It will be
implemented in Matlab environment. The main objective of this project is to develop a
newvoice recognition system that will have minimum error in identifying the speaker.
The system was built by combining a speech recognition system with a speaker
recognition system. For speech recognizer, the Dynamic Time Warping is used and for
speaker recognizer, the Mel-Frequency Cepstrum Coefficients is used. The results show
that thecombination of both algorithms has produced a robust Voice Security System.