Speaker Identification Using GMM with MFCC
Speaker identification comes under the field of digital signal processing. The earliest systems were based on acoustic phonetics built for Automatic Speech Recognition. For recognition part these systems used pattern matching and spectrum analysis. With recent advancement in technology voice recognition has become one of the efficient measure that is used to provide protection to humans computerized and electronic belongings. Voice recognition technology is the most potential technology that will make our daily lives more secure. It is one of the types of biometric that is used to identify and authenticate user on the basis of his/her voice. Voice recognition is divided into two types: text dependent and text independent. Text dependent recognition identifies user against a phrase while text independent recognition identifies the user irrespective of what he is saying. The success in both cases depends upon the various speaker characteristics which differentiate the one speaker from other. This paper targets the implementation of MFCC with GMM techniques in order to identify a speaker.
Keywords: Mel frequency Cepstral coefficient (MFCC), Gaussian Mixture Modeling, Expectation Maximization (EM) algorithm, Feature matching, Fast fourier Transform, Discrete fourier Transform, Clustering,window, Melfilter
Download Full-Text
ABOUT THE AUTHORS
Tahira Mahboob
Software Engineering, Fatima Jinnah Women University, Rawalpindi, Pakistan
Memoona Khanum
Fatima Jinnah Women University, Rawalpindi, Pakistan
Malik Sikandar Hayat Khiyal
Preston University, Islamabad
Ruqia Bibi
Fatima Jinnah Women University, Rawalpindi, Pakistan
Tahira Mahboob
Software Engineering, Fatima Jinnah Women University, Rawalpindi, Pakistan
Memoona Khanum
Fatima Jinnah Women University, Rawalpindi, Pakistan
Malik Sikandar Hayat Khiyal
Preston University, Islamabad
Ruqia Bibi
Fatima Jinnah Women University, Rawalpindi, Pakistan