Probabilistic Anchor M for Speaker Veri

نویسندگان

  • Mikaël Collet
  • Yassine Mami
چکیده

This paper presents a probabilistic approach for representing a speaker using the anchor modelling technique and discusses the relation between this new approach and the deterministic approach. In the first part, the technique of anchor modelling is presented. Then the new approach, which modelizes the various utterances of a speaker by a normal distribution in the anchor models space, is presented. A study of the relation between the two approaches (deterministic and probabilistic) leads to the definition of a combined approach. All these approaches are evaluated on the NIST 2000 database and results show improved performance of the combined approach over the deterministic and the probabilistic one.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Field test of a calling card service based on speaker verification and automatic speech recognition

In this research we have studied several human factors problems that are connected to the deployment of speaker veri cation technology in telecommunication services. We investigate the perception of the safety of a calling card service when it is protected by speaker veri cation on the 14 digit card number, and compare it to the perceived safety of speaker veri cation and PIN. Moreover, we comp...

متن کامل

A hierarchical approach to large-scale speaker recognition

This paper presents a hierarchical approach to the Large-Scale Speaker Recognition problem. In here the authors present a binary tree data-base approach for arranging the trained speaker models based on a distance measure designed for comparing two sets of distributions. The combination of this hierarchical structure and the distance measure [1] provide the means for conducting a large-scale ve...

متن کامل

Speaker verification with limited enrollment data

New methods for speaker veri cation that address the problems of limited training data and unknown telephone channel are presented. We describe a system for studying the feasibility of telephone based voice signatures for electronic documents that uses speaker veri cation with a xed test phrase but very limited data for training speaker models. We examine three methods for speaker veri cation t...

متن کامل

Audio sensors and low-complexity recognition technologies for soldier systems

The use of audio sensors for soldier systems is examined in detail. We propose two applications, speech recognition and speaker veri cation. These technologies provide two functions for a soldier system. First, the soldier has hands-free operation of his system via voice command and control. Second, identity can be veri ed for the soldier to maintain the security integrity of the soldier system...

متن کامل

Identification of Emotions from Speech using Anchor Model

A technique for refining the anchor modelling system introduced enhanced emotion detection system from speech. Anchor representation was then put on the speaker detection issue. Identification mistake transaction productivity uncovered that the anchor displaying system missed the mark supply of a state-of the-craftsmanship GMM-UBM framework. It had been more observed that its computational effi...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2005