Frame-Weighted Bayes Factor Scoring for Speaker Verification
نویسندگان
چکیده
In this paper, the Bayes factor is considered as a replacement verification criterion to the likelihood-ratio test in the context of GMM-based speaker verification. An advantage of this Bayesian method is that it allows for the incorporation of prior information and uncertainty of parameter estimates into the scoring process, complementing the Bayesian adaptation used in training. A development of Bayes factors for GMMs is presented based on incremental adaptation that is well-suited to inclusion in existing GMM-UBM systems. This method is extended to include the weighting of test frames to account for their statistical dependencies. Experiments on the 1999 NIST Speaker Recognition Evaluation corpus demonstrate improved performance over expected log-likelihood ratio scoring. These findings are supported with results from a modified version of the NIST Extended Data corpus of 2003.
منابع مشابه
Bayes factor scoring of GMMs for speaker verification
This paper implements and assesses the Bayes factor as a replacement verification criterion to the likelihood-ratio test in the context of GMM-based speaker verification. An advantage of the Bayesian method is that model parameters are considered random variables, allowing for the incorporation of prior information and uncertainty of parameter estimates into the scoring process. A novel develop...
متن کاملDiscriminative Decision Function Based Scoring Method in Joint Factor Analysis for Speaker Verification
This paper introduces a discriminative decision function scoring method for speaker recognition with the Joint Factor Analysis (JFA) system. In the scoring module of the JFA system, an approximate form of the decision function is proposed. Based on the approximation, we present a discriminative decision function by reestimating the contribution of each speech sound unit to the decision function...
متن کاملBayesian distance metric learning on i-vector for speaker verification
This thesis explores the use of Bayesian distance metric learning (Bayes dml) for the task of speaker verification using the i-vector feature representation. We propose a framework that explores the distance constraints between i-vector pairs from the same speaker and different speakers. With an approximation of the distance metric as a weighted covariance matrix of the top eigenvectors from th...
متن کاملBayesian Distance Metric Learning
This thesis explores the use of Bayesian distance metric learning (Bayes-dml) for the task of speaker verification using the i-vector feature representation. We propose a framework that explores the distance constraints between i-vector pairs from the same speaker and different speakers. With an approximation of the distance metric as a weighted covariance matrix of the top eigenvectors from th...
متن کاملSpeaker verification with multiple classifier fusion using Bayes based confidence measure
A novel framework based on Bayes-based confidence measure (BBCM) for Multiple Classifier System (MCS) fusion is proposed. As shown here, BBCM based MCS combination scheme corresponds to the ordinary Bayes fusion weighted by the reliability of each individual classifier. BBCM provides a formal model for heuristic weighting functions employed elsewhere. When compared with the ordinary Bayesian fu...
متن کامل