Asian Journal of Engineering and Technology Innovation

User Authentication Using Text - Prompted Technique

Laxmi Doddappagol, Geetha B

¹Reva Institute of Technology and Management, Rukmini Knowledge Park, Kattigenahalli, Yelahanka, Near Border Security Bustop, Bengaluru, Karnataka-560064, India.

Abstract

The authentication of the user through voice is one of the emerging technologies in today’s internet based world. Recognition of the person is process to identify the individual based on characteristic features available in the speech signals. The methods of speaker recognition can be divided into text-dependent (e.g. fixed phrase/word), text-independent (e.g. No specific phrase/word) and text- prompted. Text-prompted is one where the text on the screen is prompted and the imposter cannot know in advance the uttered phrase/word. The system gives access to the genuine users and accepts the input utterance else rejects the imposters. Hence, this technique alleviates the drawbacks of other two techniques. This project is implemented based on the techniques of text prompted. The database of 25 users is being collected. Each user utters the digits from 0-9 in English language. Various preprocessing techniques such as hamming window, normalization, end point detection and speech separation, pre emphasis are performed. The features such a Mel frequency cepstral coefficients, pitch and formant features are extracted. Support Vector Machine (SVM) is used as a classifier.The work is carried in order to provide the liveness detection of users through user voice. The accuracy of 92% to 88.7% is achieved.

Keywords:

-