Text Dependent Speaker Recognition with Back Propagation Neural Network
N K Kaphungkui1, Aditya Bihar Kandali2

1N K Kaphungkui, Department of Electronics and Communication, Dibrugarh University, Assam India.
2Dr Aditya Bihar Kandali, Department of Electrical, Jorhat Engineering College, Assam India.

Manuscript received on 18 June 2019 | Revised Manuscript received on 25 June 2019 | Manuscript published on 30 June 2019 | PP: 1431-1434 | Volume-8 Issue-5, June 2019 | Retrieval Number: E7288068519/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Speaker recognition system follows the procedure of consequently perceiving who is talking or speaking by utilizing the speaker’s particular data incorporated into the speech waveform to confirm the identity of a person. By checking the voice attributes of expression, this framework makes it conceivable to validate their character and control access to various database administrations system. This will add additional level of security to any system where security is the main concern. The primary aim of this work is to verify the speaker by extraction of speech features using MFCC and Back Propagation Neural Network as speech classifier. Voice sample of a group of four male and three female uttering the same sentence “This Voice is my password” repeatedly are collected and trained with neural network and testing the network for recognizing are done with untrained new data set with the same utterance spoken once . A specific code or target is assigned for each speakers and recognition is based on how close the network output is to the assigned code for each speaker. Recognition is based on the minimum positive error generation between the code and the actual network output. The tool for simulation is MATLAB R2013a.
Keywords: Speaker Recognition, MFCC, Text Dependent, Positive Error, BPNN, Training, Testing.

Scope of the Article: Pattern Recognition