A comparison of covariance matrix and i-vector based speaker recognition

Abstract

© Springer International Publishing AG 2017. The paper presents results of an evaluation of covariance matrix and i-vector based speaker identification methods on Serbian S70W100s120 database. Open set speaker identification evaluation scheme was adopted. The number of target speakers and the number of impostors were 20 and 60 respectively. Additional utterances from 41 speakers were used for training. Amount of data for modeling a target speaker was limited to about 4 s of speech. In this study, the i-vector base approach showed significantly better performance (equal error rate EER ~5%) than the covariance matrix based approach (EER ~16%). This small EER for the i-vector based approach was obtained after substantial reduction of the number of the parameters in universal background model, i-vector transformation matrix and Gaussian probabilistic linear discriminant analysis that is typically reported in the papers. Additionally, these experiments showed that cepstral mean and variance normalization can deteriorate EER in case of a single channel.