Symbolic Representation of Speech for Text Independent Speaker Recognition

Speaker recognition is an important branch of authenticating a speakers identity automatically based on human biological feature. We present a novel method of representing a speech by interval valued symbolic features. A method of speaker identification based on the proposed representation is also presented. It is an ideal choice for biometric which can change the future of speaker authentication mechanism as it is computationally effective and efficient. We also adopted LBZ-Vector Quantization technique for the purpose of speaker modelling using MFCC features. MFCC takes human perception sensitivity into consideration with respect to frequencies and therefore are best for speaker recognition. The technique of VQ consists of extracting a small number of representative feature vectors as an efficient means of characterizing the speaker specific features. The newly proposed model significantly reduces the dimension of feature vectors and also the time taken to classify a given speech utterances. In this work we provide a brief overview of the area of speaker recognition, describing applications and some underlying techniques. We will discuss some of the strengths and weaknesses of current speaker recognition technologies. We outline some potential future trends in research, development and applications