Fuzzy Hidden Markov Models and fuzzy NN Models in Speaker Recognition

Abstract

The fuzzy HMM algorithm is regarded as an application of the fuzzy expectation-maximization (EM) algorithm to the Baum-Welch algorithm in the HMM. The Texas Instruments p4 used speech and speaker recognition experiments and show better results for fuzzy HMMs compared with conventional HMMs. Equation and how estimation of discrete and continuous HMM parameters on based this two algorithm is explained and performance of two speech recognition method for one hundred is surveyed. This paper show better results for the fuzzy HMM, compared with the conventional HMM. After of that work we use Fuzzy-Neural Network system was proposed for Farsi speech recognition. Instead of using the fuzzy membership input with class membership desired-output during training procedure as proposed by several researches. we used the fuzzy membership input with fundamental binary desiredoutput. This can reduce the misunderstood training, decrease the training time and also improve the recognition ability.

Fuzzy Speech Recognition: A Review

“…A novel method using continuous-density hidden Markov model (CDHMM) for speech recognition based on the principle of maximizing the minimum multi-class separation margin is presented [22] i.e. large margin HMM.…”