Distributed automatic text-independent speaker identification using GMM-UBM speaker models

Chowdhury, Foezur Rahman; Selouani, Sid‐Ahmed; O’Shaughnessy, D.

doi:10.1109/ccece.2009.5090157

Cited by 10 publications

(5 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The rapid operation performance of the PSO-FFNN model has been reported since the FFNN model will be totally relying on the PSO algorithm for producing the weight coefficients without the need of performing standalone (internal) weight generation. Eventually, the mean square error (MSE) and root mean square error (RMSE) metrics are also found to be minimal [30][31][32]. MSE and RMSE metrics imply lesser error existence on the proposed model predictions.…”

Section: Resultsmentioning

confidence: 99%

Development of High Accuracy Classifier for the Speaker Recognition System

Al_Hassani

Atilla

Aydın

2021

Applied Bionics and Biomechanics

View full text Add to dashboard Cite

Speech signal is enriched with plenty of features used for biometrical recognition and other applications like gender and emotional recognition. Channel conditions manifested by background noise and reverberation are the main challenges causing feature shifts in the test and training data. In this paper, a hybrid speaker identification model for consistent speech features and high recognition accuracy is made. Features using Mel frequency spectrum coefficients (MFCC) have been improved by incorporating a pitch frequency coefficient from speech time domain analysis. In order to enhance noise immunity, we proposed a single hidden layer feed-forward neural network (FFNN) tuned by an optimized particle swarm optimization (OPSO) algorithm. The proposed model is tested using 10-fold cross-validation over different levels of Adaptive White Gaussian Noise (AWGN) (0-50 dB). A recognition accuracy of 97.83% was obtained from the proposed model in clean voice environments. However, a noisy channel is realized with lesser impact on the proposed model as compared with other baseline classifiers such as plain-FFNN, random forest (RF), K -nearest neighbour (KNN), and support vector machine (SVM).

show abstract

Section: Resultsmentioning

confidence: 99%

Development of High Accuracy Classifier for the Speaker Recognition System

Al_Hassani

Atilla

Aydın

2021

Applied Bionics and Biomechanics

View full text Add to dashboard Cite

show abstract

“…Kmeans algorithm and LBG algorithm belong to a sort of local clustering arithmetic, and the cluster result is always not global optimum under the condition of the abundant data and the fixed-setting classification number. The parameter model trained based on the cluster result has great effect on the identifying rate of the speaker identification system [4,7]. For this reason, an ant colony algorithm combined with Genetic arithmetic is proposed in the paper to improve the identifying rate.…”

Section: A Choice Of Methodsmentioning

confidence: 99%

“…[1,3]. The voiceprint recognition application of Gaussian Mixture Model (GMM) is becoming more and more popular after the 1990s because of its easiness, agility, availability and high robustness [4,14]. The voiceprint recognition technology is improved in the initialization method for the speaker model parameters training based on expectation maximization (EM) in order to raise its recognition rate in the paper.…”

Section: Introductionmentioning

confidence: 99%

Initialization in speaker model training based on expectation maximization

Wang

2013

2013 6th International Congress on Image and Signal Processing (CISP)

View full text Add to dashboard Cite

The optimized speaker model is trained by many time iterative algorithm based on expectation maximization (Abbr. EM). In the process, the choice of speaker model initial value has great influence on the final recognition effect. The most common algorithms which are used to choose the initial value are K-means algorithm and LBG algorithm at present, but the two algorithms belong to a sort of local clustering arithmetic, therefore, it is difficult for them to provide the optimal initial value. For this reason, the ant colony algorithm combined with genetic arithmetic is proposed in the paper. The comparative experiment between this algorithm and K-means algorithm has been done, and the experimental results have been obtained to verify that this algorithm can bring better recognition rate than Kmeans algorithm.

show abstract

“…The hybrid model [18] has been developed recently for speaker identification, in which two techniques, i.e., Gaussian Mixture Model (GMM) [19] and k-means clustering [20], are combined. This hybrid model can derive the precise model based on GMM extracted features for the particular speaker data, and it has been discovering the speaker clusters.…”

Section: Literature Studymentioning

confidence: 99%

Sampling Based Fuzzy Speech Clustering Systems for Faster Communication with Virtual Robotics Towards Social Applications

Prasad¹,

Narasimhulu²,

KUMAR³

et al. 2023

Preprint

View full text Add to dashboard Cite

For social applications, virtual robotics technologies require the Internet of Things (IoT) and cloud services. IoT based speaker identification is required for virtual social robotics to operate in the most secure manner possible. Moreover, humans communicate using only natural speech to virtual robotics through IoT devices; Virtual robotics are the most emerging and which are required developing the speech-based identification models in the virtual cloud platforms. State-of-the-art of speech models defines the composite technique with the combination of k-means clustering and Gaussian mixture model (GMM) for speaker identification problems. Every speaker has their own annotation, acoustic, and utterance. The Gaussian mixture model (GMM) describes the modelling of features in the speech clustering system for speaker identification applications. These traditional systems are expensive for speech clustering of large datasets when considering the time and space parameters. Sampling based fuzzy-spherical k-means and sampling based fuzzy-mini-batch k-means(mbkm) are proposed to overcome the problem of scalability issue in the implementation of speech based virtual cloud based robotic systems. For demonstrating the effectiveness of the proposed fuzzy based speaker models, extensive experimental research is conducted using a variety of benchmark speech datasets.

show abstract

Distributed automatic text-independent speaker identification using GMM-UBM speaker models

Cited by 10 publications

References 2 publications

Development of High Accuracy Classifier for the Speaker Recognition System

Development of High Accuracy Classifier for the Speaker Recognition System

Initialization in speaker model training based on expectation maximization

Sampling Based Fuzzy Speech Clustering Systems for Faster Communication with Virtual Robotics Towards Social Applications

Contact Info

Product

Resources

About