User Authentication using Voice Identification and Verification

Abstract

The use of biometric in user identification has increased tremendously in past few years. The systems based on
biometrics are quite secure and safe. There are many biometric indicators that are used for user identification with their
advantages and ids-advantages. In biometric system accuracy has been an issue and to improve on accuracy the proposed
algorithms becoming very time consuming, and a person can’t wait of 15-20 min. before he/she can be authenticated. The
human voice has special features that are different in each individual. In this work basic algorithms of voice recognition
systems are compared by using the mahanbolis distance criterion and accuracy is compared. The results presents in this
paper clearly reveals that a fair accuracy is possible with basic algorithms and these algorithms in conjunction with other
biometric identifier can provide very effective system which has a large degree of accuracy and computationally fast.