Efficient human voice recongnition system
Loading...
Date
2016
Journal Title
Journal ISSN
Volume Title
Publisher
University of Management and Technology Lahore
Abstract
The usefulness of identification of a person from the uniqueness of his/her voice is increasing with growing importance of automatic information processing. This report reviews the voice characteristics and voice identification techniques which are used in recognizing people by their voices. Voice is a random signal with continuous acoustic variability’s hence a proper digital processing algorithm has been selected for voice recognition system to extract required information from the voice sample.
Voice features are the tool on the basis of which voice recognition can be done. In this report, the implementation of VRS (Voice Recognition System) is done using MATLAB. MFCC (Mel frequency cepstral coefficients) and DTW (Dynamic Time Warping) are two algorithms which are adopted for voice feature extraction and matching.
These techniques make it possible to use human voice to recognize the identity of a person and after verification enable the control access to database services, security control for confidential information areas, providing voice recognition ability in robots.
Description
Supervised by:Muhammad Asim Butt
Keywords
Processing algorithm, Mel frequency cepstral coefficients, BS THESIS