Overview
ABSTRACT
Biometrics is a modern technique in order to indisputably prove the identity of a person by using their physical or behavioral characteristics. This article presents a new approach allowing for the automatic recognition of the speaker based on a certain number of parameters extracted from the vocal signal emitted by a person. The traditional voice recognition models integrate into multigaussian models the acoustic parameters of dozens of millisecond time windows. this method is based upon the time-frequency dynamics in the perception of voice and the selected time-scale is longer than the phonetic segment studied.
Read this article from a comprehensive knowledge base, updated and supplemented with articles reviewed by scientific committees.
Read the articleINTRODUCTION
Following on from work carried out in the field of anthropometry, biometrics is a modern technique that responds to an age-old concern to prove a person's identity indisputably by using their physical or behavioral characteristics. In this paper, we propose to study a new approach to automatic speaker recognition (ASR). Our method consists in recognizing a person based on a number of parameters, extracted from the voice signal emitted by the person, and originally represented in a qualitative way, unlike state-of-the-art parameterizations.
Exclusive to subscribers. 97% yet to be discovered!
Already subscribed? Log in!
Qualitative parametrization of the time-frequency characteristics for speaker recognition
History and state of the art
Bibliography
Exclusive to subscribers. 97% yet to be discovered!
Already subscribed? Log in!
Article included in this offer
"Digital documents and content management"
(
77 articles
)
Updated and enriched with articles validated by our scientific committees
A set of exclusive tools to complement the resources
