• Publications
  • Influence
Low-Variance Multitaper MFCC Features: A Case Study in Robust Speaker Verification
TLDR
In speech and audio applications, short-term signal spectrum is often represented using mel-frequency cepstral coefficients (MFCCs) computed from a windowed discrete Fourier transform. Expand
  • 111
  • 15
  • PDF
What else is new than the hamming window? robust MFCCs for speaker recognition via multitapering
TLDR
Multitaper methods form a spectrum estimate using multiple window functions and frequency-domain averaging. Expand
  • 37
  • 7
  • PDF
I4u submission to NIST SRE 2012: a large-scale collaborative effort for noise-robust speaker verification
TLDR
I4U is a joint entry of nine research Institutes and Universities across 4 continents to NIST SRE 2012. Expand
  • 80
  • 5
  • PDF
Temporally Weighted Linear Prediction Features for Tackling Additive Noise in Speaker Verification
TLDR
Temporally weighted linear predictive features in speaker verification hold a promise for noise-robust speaker verification. Expand
  • 71
  • 5
  • PDF
Noise-Adaptive LDA: A New Approach for Speech Recognition Under Observation Uncertainty
TLDR
Reducing the speech feature dimensionality for optimal discriminance under observation uncertainty can yield significantly improved recognition performance. Expand
  • 19
  • 5
Multitaper Estimation of Frequency-Warped Cepstra With Application to Speaker Verification
TLDR
We propose approximations of the variance and bias of the estimate of each mel-frequency cepstral coefficient and propose a general estimator which also includes multitaper estimators. Expand
  • 33
  • 4
  • PDF
Quality Measure Functions for Calibration of Speaker Recognition Systems in Various Duration Conditions
TLDR
This paper investigates the effect of utterance duration to the calibration of a modern i-vector speaker recognition system with probabilistic linear discriminant analysis (PLDA) modeling. Expand
  • 62
  • 3
  • PDF
Efficient implementation of GMM based speaker verification using sorted Gaussian mixture model
TLDR
In this paper a novel GMM structure called sorted GMM is introduced which is benefited from a fast scoring capability while its memory requirement is just marginally higher than ordinary GMM. Expand
  • 16
  • 3
  • PDF
Duration mismatch compensation for i-vector based speaker recognition systems
TLDR
We demonstrate that, as utterance duration is decreased, number of detected unique phonemes and i-vector length approaches zero in a logarithmic and non-linear fashion, respectively. Expand
  • 112
  • 2
  • PDF
Particle Swarm Optimization for Sorted Adapted Gaussian Mixture Models
TLDR
We present an efficient PSO-based scheme to enhance the performance of SGMMs by means of fine-tuning the sorting function parameters. Expand
  • 24
  • 2
  • PDF