Voice Based Biometric System

Biometric property speech used for text dependent speaker recognition system.

Biometrics is an emerging field of technology using unique and measurable physical, biological, or behavioral characteristics that can be processed to identify a person. Biometric properties of human are fingerprint, iris, face and voice.
A concise definition of biometrics is “the automatic recognition of person using distinguishing traits.”
One of the biometric property, Speech is produced as a sequence of sounds. The vibration of the vocal cords, as well as the positions, shapes, and sizes of the various articulators (such as the tongue, lips, and teeth) generate the sound being produced .The characteristics of the sound vary from person to person and can be used to identify an individual. Although typically not considered as accurate as other types of biometric identification systems, a voice recognition system can be used in conjunction with other biometric systems to create a more robust recognition system
Speaker Recognition mainly involves two modules namely feature extraction and feature matching. Feature extraction is the process that extracts a small amount of data from the speaker’s voice signal that can later be used to represent that speaker. Feature matching involves the actual procedure to identify the unknown speaker by comparing the extracted features from his/her voice input with the ones that are already stored in our speech database.
In feature extraction we find the Mel Frequency Cepstrum Coefficients (MFCC), which are based on the known variation of the human ear’s critical bandwidths with frequency and Centroid from MFCC matrix results in the speaker specific database.
In feature matching we used two methods .In first method centroids from row and column are found and codebook is made. For recognition the minimum Euclidian distance between the input utterance of an unknown speaker and the stored database should less than threshold value.
In second method using neural network toolbox,back propagation network created. for recognition Regression between output of network and desired output should higher than threshold value.