The Experts below are selected from a list of 2178 Experts worldwide ranked by ideXlab platform

Srdjan Stankovic - One of the best experts on this subject based on the ideXlab platform.

  • l statistics based space spatial frequency filtering of 2d signals in Heavy Tailed Noise
    Signal Processing, 2014
    Co-Authors: Irena Orovic, Srdjan Stankovic
    Abstract:

    Most of the commonly used stationary filtering techniques, performed either in the spatial or frequency domains, fail to produce good results for noisy signals with fast varying non-stationary structures. The filtering results could be improved by using space/spatial-frequency based non-stationary filters. Hence, a robust approach to space/spatial-frequency analysis of two-dimensional noisy signals is proposed in this paper. It is based on the two-dimensional L-estimate forms of the short-time Fourier transform, the spectrogram and the S-method. The proposed space/spatial-frequency distributions are used to define the L-estimate space-varying filtering procedure. It is designed for denoising of 2D non-stationary signals affected by the strong impulsive or mixed Heavy-Tailed and Gaussian Noise. The efficiency of the proposed procedure is tested on the examples with interferogram-like images, textures and satellite images.

  • compressed sensing based robust time frequency representation for signals in Heavy Tailed Noise
    Information Sciences Signal Processing and their Applications, 2012
    Co-Authors: Srdjan Stankovic, Irena Orovic, Moeness G Amin
    Abstract:

    A compressed sensing approach for robust time-frequency analysis of signals corrupted by strong Heavy-Tailed Noise is proposed. When using traditional time-frequency distributions and the corresponding ambiguity functions, the strong and impulsive nature of the Noise introduces spurious peaks and compromises the sparse time-frequency signal reconstruction. In order to provide accurate localization of the signal power and reduce false positives, compressed sensing is applied to the robust ambiguity function based on the L-estimation approach. This enhances the sparse time-frequency trajectories that correspond to the instantaneous frequencies of signal components. Simulation examples involving non-Gaussian Noise and signals with different instantaneous frequency laws are provided to demonstrate the effectiveness of the proposed approach.

  • ISSPA - Compressed sensing based robust time-frequency representation for signals in Heavy-Tailed Noise
    2012 11th International Conference on Information Science Signal Processing and their Applications (ISSPA), 2012
    Co-Authors: Srdjan Stankovic, Irena Orovic, Moeness G Amin
    Abstract:

    A compressed sensing approach for robust time-frequency analysis of signals corrupted by strong Heavy-Tailed Noise is proposed. When using traditional time-frequency distributions and the corresponding ambiguity functions, the strong and impulsive nature of the Noise introduces spurious peaks and compromises the sparse time-frequency signal reconstruction. In order to provide accurate localization of the signal power and reduce false positives, compressed sensing is applied to the robust ambiguity function based on the L-estimation approach. This enhances the sparse time-frequency trajectories that correspond to the instantaneous frequencies of signal components. Simulation examples involving non-Gaussian Noise and signals with different instantaneous frequency laws are provided to demonstrate the effectiveness of the proposed approach.

  • A robust form of the Ambiguity function
    2012 20th Telecommunications Forum (TELFOR), 2012
    Co-Authors: Branka Jokanovic, Srdjan Stankovic, Irena Orovic, Moeness Amin
    Abstract:

    A form of the Ambiguity function appropriate for signals affected by mixture of the Gaussian and impulsive Noise is introduced. Then, it is used for calculation of the robust distributions belonging to the Cohen class. The proposed class of robust distributions may be used for time-frequency analysis and IF estimation of a wide class of signals corrupted by a Heavy Tailed Noise. The theory is justified by a set of numerical examples.

  • An Implementation of the L-Estimate Distributions for Analysis of Signals in Heavy-Tailed Noise
    IEEE Transactions on Circuits and Systems II: Express Briefs, 2011
    Co-Authors: Nikola Zaric, Nedjeljko Lekic, Srdjan Stankovic
    Abstract:

    An analog-digital system for implementation of the robust time-frequency distributions is proposed. It provides an efficient real-time analysis of nonstationary signals corrupted by Heavy-Tailed Noise. The most important part of this system is the realization of the L-estimate short-time Fourier transform that requires sorting operation. The proposed solution is designed to execute sorting operation within a single clock cycle, regardless of the number of inputs. The performance of the proposed hardware is tested on a real signal example.

Irena Orovic - One of the best experts on this subject based on the ideXlab platform.

  • l statistics based space spatial frequency filtering of 2d signals in Heavy Tailed Noise
    Signal Processing, 2014
    Co-Authors: Irena Orovic, Srdjan Stankovic
    Abstract:

    Most of the commonly used stationary filtering techniques, performed either in the spatial or frequency domains, fail to produce good results for noisy signals with fast varying non-stationary structures. The filtering results could be improved by using space/spatial-frequency based non-stationary filters. Hence, a robust approach to space/spatial-frequency analysis of two-dimensional noisy signals is proposed in this paper. It is based on the two-dimensional L-estimate forms of the short-time Fourier transform, the spectrogram and the S-method. The proposed space/spatial-frequency distributions are used to define the L-estimate space-varying filtering procedure. It is designed for denoising of 2D non-stationary signals affected by the strong impulsive or mixed Heavy-Tailed and Gaussian Noise. The efficiency of the proposed procedure is tested on the examples with interferogram-like images, textures and satellite images.

  • compressed sensing based robust time frequency representation for signals in Heavy Tailed Noise
    Information Sciences Signal Processing and their Applications, 2012
    Co-Authors: Srdjan Stankovic, Irena Orovic, Moeness G Amin
    Abstract:

    A compressed sensing approach for robust time-frequency analysis of signals corrupted by strong Heavy-Tailed Noise is proposed. When using traditional time-frequency distributions and the corresponding ambiguity functions, the strong and impulsive nature of the Noise introduces spurious peaks and compromises the sparse time-frequency signal reconstruction. In order to provide accurate localization of the signal power and reduce false positives, compressed sensing is applied to the robust ambiguity function based on the L-estimation approach. This enhances the sparse time-frequency trajectories that correspond to the instantaneous frequencies of signal components. Simulation examples involving non-Gaussian Noise and signals with different instantaneous frequency laws are provided to demonstrate the effectiveness of the proposed approach.

  • ISSPA - Compressed sensing based robust time-frequency representation for signals in Heavy-Tailed Noise
    2012 11th International Conference on Information Science Signal Processing and their Applications (ISSPA), 2012
    Co-Authors: Srdjan Stankovic, Irena Orovic, Moeness G Amin
    Abstract:

    A compressed sensing approach for robust time-frequency analysis of signals corrupted by strong Heavy-Tailed Noise is proposed. When using traditional time-frequency distributions and the corresponding ambiguity functions, the strong and impulsive nature of the Noise introduces spurious peaks and compromises the sparse time-frequency signal reconstruction. In order to provide accurate localization of the signal power and reduce false positives, compressed sensing is applied to the robust ambiguity function based on the L-estimation approach. This enhances the sparse time-frequency trajectories that correspond to the instantaneous frequencies of signal components. Simulation examples involving non-Gaussian Noise and signals with different instantaneous frequency laws are provided to demonstrate the effectiveness of the proposed approach.

  • A robust form of the Ambiguity function
    2012 20th Telecommunications Forum (TELFOR), 2012
    Co-Authors: Branka Jokanovic, Srdjan Stankovic, Irena Orovic, Moeness Amin
    Abstract:

    A form of the Ambiguity function appropriate for signals affected by mixture of the Gaussian and impulsive Noise is introduced. Then, it is used for calculation of the robust distributions belonging to the Cohen class. The proposed class of robust distributions may be used for time-frequency analysis and IF estimation of a wide class of signals corrupted by a Heavy Tailed Noise. The theory is justified by a set of numerical examples.

Ananthram Swami - One of the best experts on this subject based on the ideXlab platform.

  • On some detection and estimation problems in Heavy-Tailed Noise
    Signal Processing, 2002
    Co-Authors: Ananthram Swami, Brian M. Sadler
    Abstract:

    We consider the problem of estimating the parameters of a linear process with stable innovations; the linear system may be non-causal and have mixed-phase (i.e., poles and/or zeros may be inside or outside the unit circle). We show that self-normalized fourth-order moments exist, can be consistently estimated, and lead to consistent estimates of the ARMA model parameters. In the context of estimating the parameters of finite variance (or deterministic) signals observed in alpha-stable Noise, we show that conventional correlation-based schemes can be used provided that the noisy data have been pre-processed by passing them through a generic zero-memory non-linearity which serves to clip the Noise. We demonstrate this through applications to harmonic retrieval and direction of arrival estimation. The idea is also seen to be useful in estimating the underlying correlation matrix of a sub-Gaussian stable process. These pre-processing ideas are also shown to be useful in the context of detection and classification, and to outperform standard approaches. Numerical results related to the Fisher information and Cramer-Rao bounds are also presented.

  • non gaussian mixture models for detection and estimation in Heavy Tailed Noise
    International Conference on Acoustics Speech and Signal Processing, 2000
    Co-Authors: Ananthram Swami
    Abstract:

    Scale mixtures of the Gaussian have been used to approximate the PDF of symmetric alpha stable processes. Such mixtures, however, cannot easily capture the Heavy-tails. We propose to use Cauchy-Gaussian mixtures which are natural in this setting. Variations of standard EM algorithms can be used to estimate the parameters of the Noise PDFs under various scenarios (Noise-only data, weak-signal assumption, partially known-signal case). The fitted mixture models can be used for detection and estimation. In the multivariate case, we present several results on Gaussian mixture approximations of sub-Gaussian PDFs, including robust estimation of the underlying correlation matrix.

  • ICASSP - Non-Gaussian mixture models for detection and estimation in Heavy-Tailed Noise
    2000 IEEE International Conference on Acoustics Speech and Signal Processing. Proceedings (Cat. No.00CH37100), 1
    Co-Authors: Ananthram Swami
    Abstract:

    Scale mixtures of the Gaussian have been used to approximate the PDF of symmetric alpha stable processes. Such mixtures, however, cannot easily capture the Heavy-tails. We propose to use Cauchy-Gaussian mixtures which are natural in this setting. Variations of standard EM algorithms can be used to estimate the parameters of the Noise PDFs under various scenarios (Noise-only data, weak-signal assumption, partially known-signal case). The fitted mixture models can be used for detection and estimation. In the multivariate case, we present several results on Gaussian mixture approximations of sub-Gaussian PDFs, including robust estimation of the underlying correlation matrix.

Pradeep Ravikumar - One of the best experts on this subject based on the ideXlab platform.

  • A Unified Approach to Robust Mean Estimation.
    arXiv: Machine Learning, 2019
    Co-Authors: Adarsh Prasad, Sivaraman Balakrishnan, Pradeep Ravikumar
    Abstract:

    In this paper, we develop connections between two seemingly disparate, but central, models in robust statistics: Huber's epsilon-contamination model and the Heavy-Tailed Noise model. We provide conditions under which this connection provides near-statistically-optimal estimators. Building on this connection, we provide a simple variant of recent computationally-efficient algorithms for mean estimation in Huber's model, which given our connection entails that the same efficient sample-pruning based estimators is simultaneously robust to Heavy-Tailed Noise and Huber contamination. Furthermore, we complement our efficient algorithms with statistically-optimal albeit computationally intractable estimators, which are simultaneously optimally robust in both models. We study the empirical performance of our proposed estimators on synthetic datasets, and find that our methods convincingly outperform a variety of practical baselines.

L Jubisa Stankovic - One of the best experts on this subject based on the ideXlab platform.

  • Analysis of polynomial FM signals corrupted by Heavy-Tailed Noise
    Signal Processing, 2004
    Co-Authors: Braham Barkat, L Jubisa Stankovic
    Abstract:

    In this paper, we consider the analysis of polynomial FM signals corrupted by additive Heavy-Tailed Noise. Standard time-frequency techniques fail to analyze such signals. For that, we propose here a new technique, named the robust polynomial Wigner-Ville distribution (r-PWVD) to handle this case. We show that this representation outperforms the robust Wigner-Ville distribution (r-WVD) and the robust spectrogram in terms of artifacts suppression and high time-frequency resolution for this class of signals. Also, we show that the peak of the r-PWVD is an accurate instantaneous frequency estimator. Examples and Monte-Carlo simulations are presented in order to validate and prove the performance of the proposed algorithm.

  • ICASSP - The robust Wigner distribution
    2000 IEEE International Conference on Acoustics Speech and Signal Processing. Proceedings (Cat. No.00CH37100), 2000
    Co-Authors: L Jubisa Stankovic, Igor Djurović, Srdjan Stanković
    Abstract:

    The standard short-time Fourier transform and the Wigner distribution can be obtained as solutions of the minimization problem, with the absolute square error as a loss function. It has been shown that some other loss functions, like for example the absolute error, can produce more robust results in the case of signals corrupted with impulse, Heavy-Tailed, Noise. This paper presents robust time frequency-signal analysis of nonstationary signals, corrupted with Heavy-Tailed Noise. For this purpose the robust Wigner distribution is introduced, as an extension of the robust M-periodogram concept. The theory is illustrated on several examples, including application of the proposed distribution on the instantaneous frequency estimation.