Fuzzy Estimation of Priors in Speaker Recognition
نویسندگان
چکیده
This paper proposes a method to estimate the a priori probability for speakers based on the training data set, speaker models and a fuzzy estimation technique. Speaker identification experiments performed on 138 Gaussian mixture speaker models in the YOHO database using the priors estimated by the fuzzy estimation method showed lower error rates than using those estimated by the probabilistic estimation method.
منابع مشابه
Speaker Adaptation in Continuous Speech Recognition Using MLLR-Based MAP Estimation
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results ar...
متن کاملSpeaker Adaptation in Continuous Speech Recognition Using MLLR-Based MAP Estimation
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results ar...
متن کاملHidden Markov models using fuzzy estimation
In the conventional hidden Markov model, the model parameters are reestimated by an iterative procedure known as the Baum-Welch method. This paper proposes an alternative procedure using fuzzy estimation, which is generalised from the fuzzy c-means and the BaumWelch methods. An extension of this approach, which uses a garbage state to deal with outlier data is also proposed. Experiments using t...
متن کاملOn-line Bayesian speaker adaptation using tree-structured transformation and robust priors
This paper presents new results by using our recently proposed on-line Bayesian learning approach for affine transformation parameter estimation in speaker adaptation. The on-line Bayesian learning technique allows updating parameter estimates after each utterance and i t can accommodate flexible forms of transformation functions as well as prior probability density function. We show through ex...
متن کاملFuzzy Evolutionary Programming for Hidden Markov Modelling in Speaker Identification
Gibbs distribution is used to represent fuzzy codebooks of individual speakers. The method of fuzzy evolutionary programming is employed to create the fuzzy codebooks and also to train hidden Markov models of speakers. This method increases the chance of attaining global maxima in Baum-Welch algorithm for hidden Markov model re-estimation. The experiments show the results of speaker identificat...
متن کامل