Performance Comparison of Multiple Speech Features for Speaker Recognition using Artifical Neural Network

Rozario, Mary Saniya; Abraham, Thomas; Mathew, Dominic

doi:10.1109/icacc48162.2019.8986182

Cited by 7 publications

(8 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The modeling and decision making in speaker recognition also vary, starting from the widely used HMM [4,16], vector quantization (VQ) [9], SVM [14], the classical technique Gaussian Mixture Model (GMM) [10,12,13], and ANN [17,18]. However, from the decision-making model, HMM is more suitable for modeling speaker features, especially in text-dependent speaker recognition systems, where there are two types of HMM, Ergodic and left-right [19].…”

Section: Introductionmentioning

confidence: 99%

Wavelet Detail Coefficient as a Novel Wavelet-MFCC Features in Text-Dependent Speaker Recognition System

Hidayat

Tajuddin

Yusuf³

et al. 2022

IIUMEJ

View full text Add to dashboard Cite

Speaker recognition is the process of recognizing a speaker from his speech. This can be used in many aspects of life, such as taking access remotely to a personal device, securing access to voice control, and doing a forensic investigation. In speaker recognition, extracting features from the speech is the most critical process. The features are used to represent the speech as unique features to distinguish speech samples from one another. In this research, we proposed the use of a combination of Wavelet and Mel Frequency Cepstral Coefficient (MFCC), Wavelet-MFCC, as feature extraction methods, and Hidden Markov Model (HMM) as classification. The speech signal is first extracted using Wavelet into one level of decomposition, then only the sub-band detail coefficient is used as the feature for further extraction using MFCC. The modeled system was applied in 300 speech datasets of 30 speakers uttering “HADIR” in the Indonesian language. K-fold cross-validation is implemented with five folds. As much as 80% of the data were trained for each fold, while the rest was used as testing data. Based on the testing, the system's accuracy using the combination of Wavelet-MFCC obtained is 96.67%. ABSTRAK: Pengecaman penutur adalah proses mengenali penutur dari ucapannya yang dapat digunakan dalam banyak aspek kehidupan, seperti mengambil akses dari jauh ke peranti peribadi, mendapat kawalan ke atas akses suara, dan melakukan penyelidikan forensik. Ciri-ciri khas dari ucapan merupakan proses paling kritikal dalam pengecaman penutur. Ciri-ciri ini digunakan bagi mengenali ciri unik yang terdapat pada sesebuah ucapan dalam membezakan satu sama lain. Penyelidikan ini mencadangkan penggunaan kombinasi Wavelet dan Mel Frekuensi Pekali Cepstral (MFCC), Wavelet-MFCC, sebagai kaedah ekstrak ciri-ciri penutur, dan Model Markov Tersembunyi (HMM) sebagai pengelasan. Isyarat penuturan pada awalnya diekstrak menggunakan Wavelet menjadi satu tahap penguraian, kemudian hanya pekali perincian sub-jalur digunakan bagi pengekstrakan ciri-ciri berikutnya menggunakan MFCC. Model ini diterapkan kepada 300 kumpulan data ucapan daripada 30 penutur yang mengucapkan kata "HADIR" dalam bahasa Indonesia. Pengesahan silang K-lipat dilaksanakan dengan 5 lipatan. Sebanyak 80% data telah dilatih bagi setiap lipatan, sementara selebihnya digunakan sebagai data ujian. Berdasarkan ujian ini, ketepatan sistem yang menggunakan kombinasi Wavelet-MFCC memperolehi 96.67%.

show abstract

Section: Introductionmentioning

confidence: 99%

Wavelet Detail Coefficient as a Novel Wavelet-MFCC Features in Text-Dependent Speaker Recognition System

Hidayat

Tajuddin

Yusuf³

et al. 2022

IIUMEJ

View full text Add to dashboard Cite

show abstract

“…In recent years, with the progress of computer technology, machine learning technology has also been rapidly developed. Neural networks as the main approach of machine learning techniques have been extensively studied and applied [17][18][19]. These machine learning methods perform well in dealing with optimisation problems.…”

Section: Introductionmentioning

confidence: 99%

High‐efficient optimisation method of antenna array radiation pattern synthesis based on multi‐layer perceptron network

Zhang

Huang

Niu

et al. 2022

IET Microwaves Antenna & Prop

View full text Add to dashboard Cite

With the increasing complexity of the application environment of antenna arrays and the increasing requirements for antenna array radiation patterns, efficient optimisation methods of antenna array radiation pattern synthesis have always been the challenges of the research. This paper proposes a method of utilising a multi‐layer perceptron network to optimise the radiation patterns of antenna array. The optimisation requirements of the antenna array radiation pattern are described by defining the objective mask function, which is designed to shape the target radiation pattern by desired features. Simulation experiments have been designed with a variety of antenna array layouts, of which the radiation patterns with different requirements were successfully synthesised and optimised. The results show that the proposed method has high efficiency and satisfactory optimisation effect in dealing with the problem of complex radiation pattern requirements.

show abstract

“…Related to this field is the blind source separation or the cocktail party effect wherein a specific speaker's voice can be extracted from a group of noisy or unwanted environment [10]. Neural networks can also be trained to model speaker features as what was done in [11] and other applications [11].…”

Section: Introductionmentioning

confidence: 99%

“…In speaker modeling, background speaker and universal background modeling methods can be utilized. In this work, we adapt the constrained maximum likelihood linear regression (CMLLR) [11] and support vector machines (SVM) methods for recognizing a person's voice. Gaussian mixture models (GMM) models the speaker and its background.…”

Section: Introductionmentioning

confidence: 99%

“…The major contributions of this work are summarized below. Background modeling employs several speakers to cover the space of the alternative hypothesis [11], while the universal background modeling has a pool of speakers in a single utterance and then train a single model. This is the predominant approach in modern speaker recognition systems [12].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Speaker Feature Modeling Utilizing Constrained Maximum Likelihood Linear Regression and Gaussian Mixture Models

Magsino¹

2020

IJATCSE

View full text Add to dashboard Cite

This paper describes a speaker recognition system based on feature extraction utilizing the constrained maximum likelihood linear regression (CMLLR) speaker adaptation, while using Gaussian mixture models (GMM) to model the speaker and background models.For the input acoustic signals, the cepstral features are derived to highlight the differences between test and training utterances. The CLSU dataset is used to test the efficiency and performance of the proposed CMLLR, Support Vector Machine, and GMM methods for modeling the speaker's voice by characterizing the speaker features.

show abstract

Performance Comparison of Multiple Speech Features for Speaker Recognition using Artifical Neural Network

Cited by 7 publications

References 11 publications

Wavelet Detail Coefficient as a Novel Wavelet-MFCC Features in Text-Dependent Speaker Recognition System

Wavelet Detail Coefficient as a Novel Wavelet-MFCC Features in Text-Dependent Speaker Recognition System

High‐efficient optimisation method of antenna array radiation pattern synthesis based on multi‐layer perceptron network

Speaker Feature Modeling Utilizing Constrained Maximum Likelihood Linear Regression and Gaussian Mixture Models

Contact Info

Product

Resources

About