A Speaker Recognition System Using Gaussian Mixture Model, EM Algorithm and K-Means Clustering

Jadhav, Ajinkya N.; Dharwadkar, Nagaraj V.

doi:10.5815/ijmecs.2018.11.03

Cited by 11 publications

(5 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have summarized the recent progress of self‐powered flexible piezoelectric acoustic sensors and machine learning algorithms . Self‐powered devices on plastics were developed using various piezoelectric materials such as ZnO NW array, PVDF nano fiber web, and perovskite thin films .…”

Section: Resultsmentioning

confidence: 99%

“…Gaussian mixture models (GMM) demonstrated the great performance for speaker recognition without extensive data, which allowed GMM one of the best choices in modeling speech data for decades . GMM‐based algorithm was adequate in showing the performance difference between the piezoelectric acoustic device and the commercialized microphone.…”

Section: Machine Learning Algorithms For Speech Processingmentioning

confidence: 99%

See 1 more Smart Citation

Flexible Piezoelectric Acoustic Sensors and Machine Learning for Speech Processing

et al. 2019

View full text Add to dashboard Cite

Flexible piezoelectric acoustic sensors have been developed to generate multiple sound signals with high sensitivity, shifting the paradigm of future voice technologies. Speech recognition based on advanced acoustic sensors and optimized machine learning software will play an innovative interface for artificial intelligence (AI) services. Collaboration and novel approaches between both smart sensors and speech algorithms should be attempted to realize a hyperconnected society, which can offer personalized services such as biometric authentication, AI secretaries, and home appliances. Here, representative developments in speech recognition are reviewed in terms of flexible piezoelectric materials, self‐powered sensors, machine learning algorithms, and speaker recognition.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Machine Learning Algorithms For Speech Processingmentioning

confidence: 99%

Flexible Piezoelectric Acoustic Sensors and Machine Learning for Speech Processing

et al. 2019

View full text Add to dashboard Cite

show abstract

“…As a result, it was determined that in an analytical form, the conceptual model of ensuring the effectiveness of the neural network phoneme recognition process can be displayed using expressions (1)(2)(3).…”

Section: Conceptual Model Of Neural Network Phoneme Recognitionmentioning

confidence: 99%

Conceptual Principles of Neuronetwork Recognition of Phonemes in the Voise Signal of Distance Learning System Members

Tereikovska,

Tereikovskyi

2023

Science, Technology and Innovation in the Modern World

View full text Add to dashboard Cite

One of the most promising ways to improve the effectiveness of distance learning systems is the introduction of interactive educational materials based on the use of voice recognition tools. The introduction of known voice recognition tools into domestic distance learning systems is associated with significant financial costs and is complicated by the need for complex adaptation to the variability of application conditions, which explains the urgency of the task of developing models, methods and tools for recognizing voice signals adapted to the conditions of the distance learning system. The subject of the research are models, methods and means of neural network recognition of voice signals of members of the distance learning system. The purpose of the work is to develop the conceptual foundations of neural network recognition of phonemes in the voice signal of members of the distance learning system. The research methodology is based on the methods of digital signal processing, the theory of neural networks, the theory of voice recognition and involves the analysis of actual problems of recognizing members’ voice signals in the distance learning system, the analysis of modern neural network solutions in the field of voice recognition, the development of a conceptual model and the principles of effective application of neural networks for recognizing phonemes in the voice signal of distance learning system members. As a result of the conducted research, it was determined that the main functions of the module for recognizing the voice signals of members of the distance learning system are identification at the entrance, determination of the voice response in the process of computer testing, and determination of the voice command when using the services. It is substantiated that the functioning of such a module can be implemented on the basis of neural network means of phoneme recognition. At the same time, the conditions for the implementation of voice recognition tools are characterized by limitations on the development period, the involvement of labor resources, and restrictions on access to the audio recording databases necessary for the training of the neural network model, and lead to the need to forecast the load on the server of the distance learning system and take into account the requirements for the voice transmission channel signals, microphone, room acoustics, hardware and software. Inadequate adaptation of known neural network means of recognizing voice signals to the conditions of domestic distance learning systems has been proven. It is substantiated that effective recognition can be implemented on the basis of the neural network method of recognizing selected phonemes, which involves the implementation of basic procedures related to the selection of the type and parameters of the neural network model, adaptation of the learning method, effective coding of the initial parameters, formation of an effective training sample, and forecasting sufficient computing resources. A conceptual model of neural network recognition of phonemes in the voice signal of members of the distance learning system has been developed, which ensures the formalization of the process of building neural network tools for phoneme recognition. The principles of the use of neural networks for phoneme recognition have been formed, which should be used as a basis for creating models of the processes of using neural network means of phoneme recognition in the voice signal of members of the distance learning system. The proposed list of basic procedures for building effective neural network tools for phoneme recognition, the developed conceptual model of neural network recognition of phonemes in the voice signal of members of the distance learning system, and the formed principles of using neural networks for phoneme recognition form the conceptual foundations of neural network recognition of phonemes in the voice signal of members of the distance learning system.

show abstract

“…They determined that the GMM based MFCC was the optimum feature extraction technique compared to the BFCC [12]. Jadhav et al (2018), proposed a speaker recognition system using GMM. This system used the voice recordings of 28 people and achieved an accuracy rate of 96.42% [13].…”

Section: Introductionmentioning

confidence: 99%

“…Jadhav et al (2018), proposed a speaker recognition system using GMM. This system used the voice recordings of 28 people and achieved an accuracy rate of 96.42% [13].…”

Section: Introductionmentioning

confidence: 99%

Markov Model Based Real Time Speaker Recognition using K-Means, Fast Fourier Transform and Mel Frequency Cepstral Coefficients

Borandağ

2019

Celal Bayar Üniversitesi Fen Bilimleri Dergisi

View full text Add to dashboard Cite

In this study, which was carried out using a combination of machine learning and sound processing methods, a speaker recognition system and application were developed using real-time Mel Frequency Cepstral Coefficients (MFCC) features and Markov chain model classifier. A sound sample was taken from each speaker for the training of the system and these sound samples were processed in Fast Fourier Transform and MFCC feature extraction algorithms. The MFCC features were clustered using the kmeans clustering algorithm. A Markov chain model was created for each speaker by using the outputs obtained after clustering. By deducting the characteristic features of the voice of the speaker, the person who was talking in the society and how long and at which time intervals they spoke during the conversation was determined in real time with high accuracy.

show abstract

A Speaker Recognition System Using Gaussian Mixture Model, EM Algorithm and K-Means Clustering

Cited by 11 publications

References 14 publications

Flexible Piezoelectric Acoustic Sensors and Machine Learning for Speech Processing

Flexible Piezoelectric Acoustic Sensors and Machine Learning for Speech Processing

Conceptual Principles of Neuronetwork Recognition of Phonemes in the Voise Signal of Distance Learning System Members

Markov Model Based Real Time Speaker Recognition using K-Means, Fast Fourier Transform and Mel Frequency Cepstral Coefficients

Contact Info

Product

Resources

About