Utterance Based Speaker Identification Using ANN

Dipankar Das

doi:10.5121/ijcsea.2014.4403

Abstract

In this paper we present the implementation of speaker identification sy stem using artificial neural network with digital signal processing. The system is designed to work with the text -dependent speaker identification for Bangla Speech. The utterances of speakers are recorded for specific Bangla words using an audio wave recorder. The speech features are acquired by the digital signal processing technique. The identification of speaker using frequency domain data is performed using backpropagation algorithm. Hamming window and Blackman -Harris window are used to investigate bet ter speaker identification performance. Endpoint detection of speech is developed in order to achieve high accuracy of the system.

Highlights

Most of us can recognize a known person’s voice without seeing him, this ability of recognition is known as speaker identification
We propose the text-dependent speaker identification system using Artificial Neural Network (ANN)
Since the computational load of the network increases with the increasing of the hidden layer nodes, so the network takes more time to reach the error goal

Summary

Introduction

Most of us can recognize a known person’s voice without seeing him, this ability of recognition is known as speaker identification. Human’s abilities both to understand the speech and to recognize the speakers from their voices have inspired many scientists to research in this field. Prior to mid 1960’s, most of speech processing systems were based on analog hardware implementation. Since the advent of inexpensive digital computers and pulse code modulation (PCM), the speech area has undergone many significant advances. Successful speech processing systems require knowledge in many disciplines including acoustic wave spectrum, pattern recognition, and artificial intelligence techniques. Speech technology includes the following areas: speech enhancement, speaker separation, speech coding, speech recognition, speech synthesis, and speaker recognition. The area of speaker recognition can be divided into speaker identification and speaker verification. In this paper the main emphasis is on the speaker identification problem

Methods

Discussion

Conclusion