Accent Identification for Saudi Speakers

 

The goal of this project is to create an intelligent program that is able to listen to a Saudi speaker then classify him/her according to him/her accent. This project also aims at identifying the speaker’s gender and identifies the speaker.

 

Table of Contents

Section 1  Introduction    Section1

Section 2  The SAAVB    Section2

2.1 What is SAAVB?

2.2 What is the objective of SAVVB?

2.3 How the process of collecting data works?

   2.3.1 How the Telephone Interface System is built?

   2.3.2 How the Call Flow Recording System is operated?

         2.3.2.1 The Call Receive program

        2.3.2.2 The Recording Call Flow program

   2.3.3 The Transcription Setup

        2.3.3.1 The Hardware Setup

        2.3.3.2 The Software Setup

   2.3.4 The Backup and Data Management System

        2.3.4.1 The Recording Setup Backup and Management System

   2.3.5 The Transcription Backup and Data management System

Section 3    Feature Extraction    Section3

3.1  Introduction

3.2  Signal-processing front end

3.3  Mel-Frequency Cepstrum Coefficients

Section 4 Hidden Markov Model toolKit    Section4

4.1 Introduction to HMM

4.2 Markov model

4.3 Hidden Markov Model

4.4 Introduction to HTK

Section 5   Gender Identification     Section5
 

5.1  What is gender identification?

5.2  Another used algorithm: The Peak algorithm

5.3 How to use HTK for Gender Identification

     5.3.1 Training step

     5.3.2 Testing step

Section 6 Accent Identification      Section6
 Section 7 Speaker Identification   Section7
 Section 8 Developed Programs  
 Section 9 Experiments Results   Section9