Voice for Identity Recognition

Abstract

Biometric technology is fast gaining popularity as means of security measures which, helps to reduce cases of fraud and theft due to its use of physical characteristics and trains for the identification of individuals. Since the most popular methods of keeping information and resources secure are to use password and User ID/PIN protection. These schemes require the users to authenticate themselves by entering a "secret “password that they had previously created or were assigned to. These systems are prone to hacking, either from an attempt to crack the password or from passwords which were not unique. A Biometric voice Identification system is the one of the user's "voice" becomes the password/PIN. Voice characteristics of an individual are unique and therefore can be used to authenticate a user's access to various systems, the rate and pitch of each pronounced vowel is not the same. This poses the major concern for voice identification systems that is how to account for the variations in one's voice, each time the voice identification occurs. Furthermore, they tend to have a high false reject rate because of background noise and other variables. Robust reliable voice identification system (VIS) (MATLAB software) was designed in this project. The system able to identify speaker by their voice through two major phases; training or enrollment phase and identification or testing phase , in training phase, features of speech signal will extract and store as a reference features. In the identification phase the same features will extract from an input utterance of the speaker whose identity will be determined, then the decision will be taken according to the matching result between the reference template and the template derived from the input utterance. Results will display in connected circuit, green led will light in matching case and the red one accompanied with buzzer beep if not matching