Limited Data Speaker Verification: Fusion of Features

The present work demonstrates experimental evaluation of speaker verification for different speech feature extraction techniques with the constraints of limited data (less than 15 seconds). The state-of-the-art speaker verification techniques provide good performance for sufficient data (greater than 1 minutes). It is a challenging task to develop techniques which perform well for speaker verification under limited data condition. In this work different features like Mel Frequency Cepstral Coefficients (MFCC), Linear Prediction Cepstral Coefficients (LPCC), Delta (4), Delta-Delta (44), Linear Prediction Residual (LPR) and Linear Prediction Residual Phase (LPRP) are considered. The performance of individual features is studied and for better verification performance, combination of these features is attempted. A comparative study is made between Gaussian mixture model (GMM) and GMM-universal background model (GMM-UBM) through experimental evaluation. The experiments are conducted using NIST-2003 database. The experimental results show that, the combination of features provides better performance compared to the individual features. Further GMM-UBM modeling gives reduced equal error rate (EER) as compared to GMM.

Download Full-text

Multilingual Speaker Identification by Combining Evidence from LPR and Multitaper MFCC

Journal of Intelligent Systems ◽

10.1515/jisys-2013-0038 ◽

2013 ◽

Vol 22 (3) ◽

pp. 241-251 ◽

Cited By ~ 2

Author(s):

B.G. Nagaraja ◽

H.S. Jayanna

Keyword(s):

Speaker Recognition ◽

Linear Prediction ◽

Speaker Identification ◽

Gaussian Mixture ◽

Limited Data ◽

Mel Frequency Cepstral Coefficients ◽

Residual Phase ◽

Combining Evidence ◽

Prediction Residual ◽

Hilbert Envelope

AbstractIn this work, the significance of combining the evidence from multitaper mel-frequency cepstral coefficients (MFCC), linear prediction residual (LPR), and linear prediction residual phase (LPRP) features for multilingual speaker identification with the constraint of limited data condition is demonstrated. The LPR is derived from linear prediction analysis, and LPRP is obtained by dividing the LPR using its Hilbert envelope. The sine-weighted cepstrum estimators (SWCE) with six tapers are considered for multitaper MFCC feature extraction. The Gaussian mixture model–universal background model is used for modeling each speaker for different evidence. The evidence is then combined at scoring level to improve the performance. The monolingual, crosslingual, and multilingual speaker identification studies were conducted using 30 randomly selected speakers from the IITG multivariability speaker recognition database. The experimental results show that the combined evidence improves the performance by nearly 8–10% compared with individual evidence.

Download Full-text

i-Vector-Based Speaker Verification on Limited Data Using Fusion Techniques

Journal of Intelligent Systems ◽

10.1515/jisys-2017-0047 ◽

2018 ◽

Vol 29 (1) ◽

pp. 565-582

Author(s):

T.R. Jayanthi Kumari ◽

H.S. Jayanna

Keyword(s):

Linear Prediction ◽

Vocal Tract ◽

Speaker Verification ◽

Excitation Source ◽

Limited Data ◽

Score Level Fusion ◽

Verification System ◽

Modeling Techniques ◽

Prediction Residual ◽

Level Fusion

Abstract In many biometric applications, limited data speaker verification plays a significant role in practical-oriented systems to verify the speaker. The performance of the speaker verification system needs to be improved by applying suitable techniques to limited data condition. The limited data represent both train and test data duration in terms of few seconds. This article shows the importance of the speaker verification system under limited data condition using feature- and score-level fusion techniques. The baseline speaker verification system uses vocal tract features like mel-frequency cepstral coefficients, linear predictive cepstral coefficients and excitation source features like linear prediction residual and linear prediction residual phase as features along with i-vector modeling techniques using the NIST 2003 data set. In feature-level fusion, the vocal tract features are fused with excitation source features. As a result, on average, equal error rate (EER) is approximately equal to 4% compared to individual feature performance. Further in this work, two different types of score-level fusion are demonstrated. In the first case, fusing the scores of vocal tract features and excitation source features at score-level-maintaining modeling technique remains the same, which provides an average reduction approximately equal to 2% EER compared to feature-level fusion performance. In the second case, scores of the different modeling techniques are combined, which has resulted in EER reduction approximately equal to 4.5% compared with score-level fusion of different features.

Download Full-text

Comparison of feature extraction and normalization methods for speaker recognition using grid-audiovisual database

Indonesian Journal of Electrical Engineering and Computer Science ◽

10.11591/ijeecs.v18.i2.pp782-789 ◽

2020 ◽

Vol 18 (2) ◽

pp. 782

Author(s):

Musab T. S. Al-Kaltakchi ◽

Haithem Abd Al-Raheem Taha ◽

Mohanad Abd Shehab ◽

Mohamed A.M. Abdullah

Keyword(s):

Feature Extraction ◽

Speaker Recognition ◽

Speaker Identification ◽

Gaussian Mixture ◽

Identification Accuracy ◽

Identification System ◽

Good Representation ◽

Mel Frequency Cepstral Coefficients ◽

Normalization Methods ◽

Cepstral Coefficients

In this paper, different feature extraction and feature normalization methods are investigated for speaker recognition. With a view to give a good representation of acoustic speech signals, Power Normalized Cepstral Coefficients (PNCCs) and Mel Frequency Cepstral Coefficients (MFCCs) are employed for feature extraction. Then, to mitigate the effect of linear channel, Cepstral Mean-Variance Normalization (CMVN) and feature warping are utilized. The current paper investigates Text-independent speaker identification system by using 16 coefficients from both the MFCCs and PNCCs features. Eight different speakers are selected from the GRID-Audiovisual database with two females and six males. The speakers are modeled using the coupling between the Universal Background Model and Gaussian Mixture Models (GMM-UBM) in order to get a fast scoring technique and better performance. The system shows 100% in terms of speaker identification accuracy. The results illustrated that PNCCs features have better performance compared to the MFCCs features to identify females compared to male speakers. Furthermore, feature wrapping reported better performance compared to the CMVN method.

Download Full-text