The Experts below are selected from a list of 4404 Experts worldwide ranked by ideXlab platform

Biing-hwang Juang - One of the best experts on this subject based on the ideXlab platform.

  • Nonlinear Compensation Using the Gauss–Newton Method for Noise-Robust Speech Recognition
    IEEE Transactions on Audio Speech and Language Processing, 2012
    Co-Authors: Yong Zhao, Biing-hwang Juang
    Abstract:

    In this paper, we present the Gauss-Newton Method as a unified approach to estimating noise parameters of the prevalent nonlinear compensation models, such as vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT), for noise-robust speech recognition. While iterative estimation of noise means in a generalized EM framework has been widely known, we demonstrate that such approaches are variants of the Gauss-Newton Method. Furthermore, we propose a novel noise variance estimation algorithm that is consistent with the Gauss-Newton principle. The formulation of the Gauss-Newton Method reduces the noise estimation problem to determining the Jacobians of the corrupted speech parameters. For sampling-based compensations, we present two Methods, sample Jacobian average (SJA) and cross-covariance (XCOV), to evaluate these Jacobians. The proposed noise estimation algorithm is evaluated for various compensation models on two tasks. The first is to fit a Gaussian mixture model (GMM) model to artificially corrupted samples, and the second is to perform speech recognition on the Aurora 2 database. The significant performance improvements confirm the efficacy of the Gauss-Newton Method to estimating the noise parameters of the nonlinear compensation models.

  • Non-linear noise compensation for robust speech recognition using Gauss-Newton Method
    2011 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP), 2011
    Co-Authors: Yong Zhao, Biing-hwang Juang
    Abstract:

    In this paper, we present the Gauss-Newton Method as a unified approach to optimizing non-linear noise compensation models, such as vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT). We demonstrate that the commonly used approaches that iteratively approximate the noise parameters in an EM framework are variants of the Gauss-Newton Method. Through the formulation of the Gauss-Newton Method for estimating noise means and variances, the noise estimation problems are reduced to determining the Jacobians of the noisy speech distributions. For the sampling-based compensations, we present two Methods, sample Jacobian average (SJA) and cross-covariance (XCOV), to evaluate the Jacobians. Experiments on the Aurora 2 database verify the efficacy of the Gauss-Newton Method to these noise compensation models.

  • ICASSP - Non-linear noise compensation for robust speech recognition using Gauss-Newton Method
    2011 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP), 2011
    Co-Authors: Yong Zhao, Biing-hwang Juang
    Abstract:

    In this paper, we present the Gauss-Newton Method as a unified approach to optimizing non-linear noise compensation models, such as vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT). We demonstrate that the commonly used approaches that iteratively approximate the noise parameters in an EM framework are variants of the Gauss-Newton Method. Through the formulation of the Gauss-Newton Method for estimating noise means and variances, the noise estimation problems are reduced to determining the Jacobians of the noisy speech distributions. For the sampling-based compensations, we present two Methods, sample Jacobian average (SJA) and cross-covariance (XCOV), to evaluate the Jacobians. Experiments on the Aurora 2 database verify the efficacy of the Gauss-Newton Method to these noise compensation models.

Yong Zhao - One of the best experts on this subject based on the ideXlab platform.

  • Nonlinear Compensation Using the Gauss–Newton Method for Noise-Robust Speech Recognition
    IEEE Transactions on Audio Speech and Language Processing, 2012
    Co-Authors: Yong Zhao, Biing-hwang Juang
    Abstract:

    In this paper, we present the Gauss-Newton Method as a unified approach to estimating noise parameters of the prevalent nonlinear compensation models, such as vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT), for noise-robust speech recognition. While iterative estimation of noise means in a generalized EM framework has been widely known, we demonstrate that such approaches are variants of the Gauss-Newton Method. Furthermore, we propose a novel noise variance estimation algorithm that is consistent with the Gauss-Newton principle. The formulation of the Gauss-Newton Method reduces the noise estimation problem to determining the Jacobians of the corrupted speech parameters. For sampling-based compensations, we present two Methods, sample Jacobian average (SJA) and cross-covariance (XCOV), to evaluate these Jacobians. The proposed noise estimation algorithm is evaluated for various compensation models on two tasks. The first is to fit a Gaussian mixture model (GMM) model to artificially corrupted samples, and the second is to perform speech recognition on the Aurora 2 database. The significant performance improvements confirm the efficacy of the Gauss-Newton Method to estimating the noise parameters of the nonlinear compensation models.

  • Non-linear noise compensation for robust speech recognition using Gauss-Newton Method
    2011 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP), 2011
    Co-Authors: Yong Zhao, Biing-hwang Juang
    Abstract:

    In this paper, we present the Gauss-Newton Method as a unified approach to optimizing non-linear noise compensation models, such as vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT). We demonstrate that the commonly used approaches that iteratively approximate the noise parameters in an EM framework are variants of the Gauss-Newton Method. Through the formulation of the Gauss-Newton Method for estimating noise means and variances, the noise estimation problems are reduced to determining the Jacobians of the noisy speech distributions. For the sampling-based compensations, we present two Methods, sample Jacobian average (SJA) and cross-covariance (XCOV), to evaluate the Jacobians. Experiments on the Aurora 2 database verify the efficacy of the Gauss-Newton Method to these noise compensation models.

  • ICASSP - Non-linear noise compensation for robust speech recognition using Gauss-Newton Method
    2011 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP), 2011
    Co-Authors: Yong Zhao, Biing-hwang Juang
    Abstract:

    In this paper, we present the Gauss-Newton Method as a unified approach to optimizing non-linear noise compensation models, such as vector Taylor series (VTS), data-driven parallel model combination (DPMC), and unscented transform (UT). We demonstrate that the commonly used approaches that iteratively approximate the noise parameters in an EM framework are variants of the Gauss-Newton Method. Through the formulation of the Gauss-Newton Method for estimating noise means and variances, the noise estimation problems are reduced to determining the Jacobians of the noisy speech distributions. For the sampling-based compensations, we present two Methods, sample Jacobian average (SJA) and cross-covariance (XCOV), to evaluate the Jacobians. Experiments on the Aurora 2 database verify the efficacy of the Gauss-Newton Method to these noise compensation models.

Paulo Roberto De Oliveira - One of the best experts on this subject based on the ideXlab platform.

Ioannis K. Argyros - One of the best experts on this subject based on the ideXlab platform.

Arye Nehorai - One of the best experts on this subject based on the ideXlab platform.

  • Information-Driven Distributed Maximum Likelihood Estimation Based on Gauss-Newton Method in Wireless Sensor Networks
    IEEE Transactions on Signal Processing, 2007
    Co-Authors: Tong Zhao, Arye Nehorai
    Abstract:

    In this paper, we develop an energy-efficient distributed estimation Method that can be used in applications such as the estimation of a diffusive source and the localization and tracking of an acoustic target in wireless sensor networks. We first propose a statistical measurement model in which we separate the linear and nonlinear parameters. This modeling strategy reduce the processing complexity. We then study the distributed implementation of the Gauss-Newton Method in the maximum likelihood estimation. After that we propose a fully distributed estimation approach based on an incremental realization of the Gauss-Newton Method. We derive three modifications of the basic algorithm to improve the distributed processing performance while still considering the energy restriction. We implement the idea of information-driven collaborative signal processing and provide a sensor-node scheduling scheme in which the Cramer-Rao bound (CRB) is used as the performance and information utility measure to select the next sensor node. Numerical examples are used to study the performance of the distributed estimation, and we show that of the Methods considered here, the proposed multiple iteration Kalman filtering Method has the most advantages for wireless sensor networks.