An MFCC-based text-independent speaker identification system for access control

摘要

In recent years, by merit of convenient and unique features, bio‐authentication techniques havernbeen applied to identify and authenticate a person based on his/her spoken words and/orrnsentences. Among these techniques, speaker recognition/identification is the most convenientrnone, providing a secure and strong authentication solution viable for a wide range of applications.rnIn this paper, to safeguard real‐world objects, like buildings, we develop a speaker identificationrnsystem named mel frequency cepstral coefficients (MFCC)‐based speaker identification systemrnfor access control (MSIAC for short), which identifies a speaker U by first collecting U's voice signalsrnand converting the signals to frequency domain. An MFCC‐based human auditory filteringrnmodel is utilized to adjust the energy levels of different frequencies as U's voice quantified features.rnNext, a Gaussian mixture model is employed to represent the distribution of the logarithmicrnfeatures as U's specific acoustic model. When a person, eg, x, would like to access a real‐worldrnobject protected by the MSIAC, x's acoustic model is compared with known‐people's acousticrnmodels. Based on the identification result, the MSIAC will determine whether the access willrnbe accepted or denied.