Speaker Recognition Application using MFCC GUI Concept

Speech, iris, face, finger print are the fundamental parameters that can help in designing a biometric authentication system. These kinds of systems are helpful in recognizing the identity of the authenticated person. The voice is a signal of infinite information. Speech based biometric authentication system is the most successful one due to simple nature of acquiring the voice and uniqueness associated with it. In this paper Speech processing is emerged as one of the important application area of digital signal processing. Various fields for research in speech processing are speech recognition, speaker recognition, speech synthesis, speech coding etc. The objective of automatic speaker recognition is to extract, characterize and recognize the information about speaker identity. The mel frequency cepstral coefficient (MFCC) is one of the most important features required among various kinds of speech applications. Fundamental motto is to design a safety box which is being operated by the voice through MATLAB - MFCC and GUI. The system developed is able to recognize specific user by extracting various characteristics of speech signals. Tested on various sample speech which belongs to male and female of various age group. The level of accuracy obtained from MFCC is much higher than of the other concepts like Dynamic Time Wrapping and Perceptual Linear Prediction. Another advantage is its ease of usage which is being obtained through its GUI concept.