Speaker recognition is the process of automatically recognizing who is speaking [4][24] on the basis of information obtained from the speech waves[13],[18],[20],[21][22]. It should be noted that this process of recognition is different from speech recognition which is not [5][16] biometrics and it’s defined as the process of recognizing what is being said, e.g., dictation of words by an individual for computer understanding (speech to text recognition). The goal of speech recognition is to answer the question “what are you saying?” This means that speech recognition is not interested in who is speaking. Whenever a speech recognition is used as part of an authentication process, a spoken version of traditional typed …show more content…
This is a yes or no decision which is a 1:1 classification[10]. Speaker Identification in other words is the use of database of voices to identify an individual based on answering the question “Who are you?” It is also assumed that the unknown individual voice must come from a fixed set of known database of speakers. No Identity claim is made by the speaker where the system must perform a 1:N classification [10] i.e. Matching of the unknown voice print/model to a database of many known voice prints/ …show more content…
A Brief History of Speaker Recognition
Speaker recognition has come a long way after over a decade of speech recognition. As discussed by Nilu Singh [1] that in early 1960 Lawrence Kersta of Bell Laboratories was the first person to develop a model for acoustic speech production which is based on “spectrographic voice verification”. Since the shape and size of vocal tract vary from one speaker to another it shows the differences in resonance frequencies [2] and [1] but now a days most of the speaker recognition system is based on the spectral information.
The Swedish Professor named by Gunnar Fant developed the first acoustic speech model physiological components in 1960 after his analysis with x-rays on people making a particular phonic sound. This was later expanded upon by Dr. Joseph Perkell in 1970 [2] [7] who provided a more detailed explanations and findings of the complex behavioral components of speech.
Texas Instruments developed [2] the first prototype of speaker recognition is developed in 1976, and was tested by the American Air Force and the MITRE Corporation In the mid 1980s the NIST(National Institute of Standards and Technology) [7] Speech Group in the United States of America is established to promote and study the use of speech processing