A fuzzy-GMM classifier for multilingual speaker identification

Abstract

In this paper,a new modeling approach is proposed by hybriding the features of expectation-maximization algorithm(GMM) and fuzzy c-means algorithm(FCM). Based on the analysis over conventional GMM technique, we suggested a new speaker identification system by fusing GMM (optimized using EM algorithm) and FCM, to improve the identification rate further in multilingual speaker identification system. The proposed technique and GMM technique was evaluated in mono and multilingual environments. Experiments were done also by varying the initial code books for generating speaker model. The experimental result shows improvements on a combined FGMM system, which employs fusion for the multilingual context with varying initial code books gives an improvement of minimum 2.98% than existing GMM approach. MFCC technique is used for extracting the features. The algorithms were compared using TIMIT database of 54 speakers speaking 3 languages like English, Hindi and Tamil.

Accurate Reader Identification for the Arabic Holy Quran Recitations Based on an Enhanced VQ Algorithm

“…Moreover, the majority of voice signal processing was limited to the English language. There was scarce work specifically in processing voice signals for the Arabic language and the Holy Quran [11]. Consequently, with all the technological development, these problems must be solved to amplify the number of learners.…”

Phoneme analysis for multiple languages with fuzzy‐based speaker identification

“…However, other methods such as vector quantisation [17,18] have their spots. A little research is made for fuzzy classification [8,[19][20][21][22], but most are quite dubious when describing their methods for both models and data. Furthermore, most recent research has converged to Neural Network variants, such as Deep Neural Networks [11,23,24], Convolutional Neural Networks [25], and others [26,27].…”

“…The literature presents a wide range of experiments for this biometry. To represent a speaker, the Mel-Frequency Cepstrum Coefficients (MFCC) are widely adopted [7][8][9][10] as voice-print, even though the state of the art has shifted from it to i-vectors [9,11] and then towards x-vectors [12]. Also, there are some variants for those representations [13] and few using fuzzy information theory [14,15].…”

Investigating the use of multiple languages for crisp and fuzzy speaker identification

“…The literature has a great variety for this biometry. When representing a speaker, the Mel-Frequency Cepstrum Coeffi-cients (MFCC) and some variations are still widely adopted [5,6,7,8,9,10,11], even though the state-of-the-art has shifted from it to i-vectors [10,12] and then towards x-vectors [13].…”

“…However, other methods such as vector quantisation (clustering) [15,16] have their spots. Little research is made for fuzzy classification [17,18,8,19,20], but the majority is quite dubious when describing their methods for both models and data. Furthermore, most recent research has converged to Neural Network variants, such as Deep Neural Networks [21,12,5], Convolutional Neural Networks [22], and others [23,24].…”