Linear Prediction Error Research Articles

Increasing anti-spoofing protection in speaker verification using linear prediction

This article addresses the problem of anti-spoofing protection in an automatic speaker verification (ASV) system. An improved version of a previously proposed spoofing countermeasure is presented. The presented method is based on the analysis of linear prediction error that results from both short- and long-term prediction of the input speech signal. It was observed that non-natural speech signals, i.e., synthetic or converted speech, were predicted in a different way than genuine speech. Therefore, in contrast to the classical linear prediction analysis, where usually only the prediction coefficients are analyzed, in the proposed approach the residual (error) signals were examined. During this analysis, 23 various prediction parameters were extracted, such as the energy of the prediction error, prediction gains and temporal parameters related to the prediction error signals. Various binary classifiers were researched to separate human and spoof classes, however the support vector machines with radial basis function (SVM-RBF) yielded the best results. When tested on the corpora provided for the ASVspoof 2015 Challenge, the proposed countermeasure returned better results than the previous version of the algorithm and, in most of the cases, the baseline spoofing detector based on the local binary patterns (LBP). It is hoped that the proposed method can be part of a generalized spoofing countermeasure helping to increase security of ASV systems.

Motion prediction in MRI-guided radiotherapy based on interleaved orthogonal cine-MRI

In-room cine-MRI guidance can provide non-invasive target localization during radiotherapy treatment. However, in order to cope with finite imaging frequency and system latencies between target localization and dose delivery, tumour motion prediction is required. This work proposes a framework for motion prediction dedicated to cine-MRI guidance, aiming at quantifying the geometric uncertainties introduced by this process for both tumour tracking and beam gating. The tumour position, identified through scale invariant features detected in cine-MRI slices, is estimated at high-frequency (25 Hz) using three independent predictors, one for each anatomical coordinate. Linear extrapolation, auto-regressive and support vector machine algorithms are compared against systems that use no prediction or surrogate-based motion estimation. Geometric uncertainties are reported as a function of image acquisition period and system latency. Average results show that the tracking error RMS can be decreased down to a [0.2; 1.2] mm range, for acquisition periods between 250 and 750 ms and system latencies between 50 and 300 ms. Except for the linear extrapolator, tracking and gating prediction errors were, on average, lower than those measured for surrogate-based motion estimation. This finding suggests that cine-MRI guidance, combined with appropriate prediction algorithms, could relevantly decrease geometric uncertainties in motion compensated treatments.

Blind multipath time varying channel estimation using recursive Cholesky update

This paper proposes a simple and high performance least square solver using recursive update of modified Cholesky factors. While most of the methods require inversion of correlation matrix, the proposed method is based on recursive updating of the modified Cholesky factors of correlation matrix and linear prediction error to estimate time varying channel coefficient, without finding the correlation matrix. Blind channel estimation of time varying channel is considered for a SIMO system using only second order statistics (SoS) of the data. The time varying channel is described by complex exponential basis expansion model (CE-BEM). The performance of the proposed method is evaluated by simulation results to demonstrate its effectiveness.

Discriminative structured dictionary learning with hierarchical group sparsity

Learning adaptive dictionaries for sparse coding has been the focus of latest research as it provides a promising way to maximize the efficiency of sparse representation. In particular, learning discriminative dictionaries rather than reconstructive ones has demonstrated significantly improved performance in pattern recognition. In this paper, a powerful method is proposed for discriminative dictionary learning. During the dictionary learning process, we enhance the discriminability of sparse codes by promoting hierarchical group sparsity and reducing linear prediction error on sparse codes. With the employment of joint within-class collaborative hierarchical sparsity, our method is able to learn adaptive dictionaries from labeled data for classification, which encourage coefficients to be sparse at both group level and singleton level and thus enforce the separability of sparse codes. Benefiting from joint dictionary and classifier learning, the discriminability of sparse codes is further strengthened. An efficient alternating iterative scheme is presented to solve the proposed model. We applied our method to face recognition, object recognition and scene classification. Experimental results have demonstrated the excellent performance of our method in comparison with existing discriminative dictionary learning approaches.

Autoregressive Hidden Markov Model and the Speech Signal

Abstract This paper introduces an autoregressive hidden Markov model (HMM) and demonstrates its application to the speech signal. In this variant of the HMM the observed signal is assumed to be Gaussian autoregressive and the probability density function is derived based on an approximation of the linear prediction error. A Baum-Welch style set of re-estimation formulas are then derived and used to infer the model parameters for a given data set, which correspond to linguistic structure in the context of speech data. The new set of re-estimation formulas are then applied to speech data and experimental results demonstrate inference of broad phonetic categories without prior knowledge of linguistic information. The experimental results and stability of this model are then briefly contrasted with historic experiments wherein phonetic information has been inferred directly from the speech signal using a similar autoregressive model.

Adaptive Radar Sensitivity Time Control Based on Linear Prediction Sea Clutter

Sea clutter is interference background in radar target processing. It is difficult to detect long-range weak target in sea clutter. Sea clutter signal changes complex, with high intensity. For close signal, radar receiver will overload and reach saturation point. We are interested in the most dynamic range for displaying target echo and fast changing component of clutter. By analyzing spatial and temporal correlation of sea clutter, slow changing component can be minus from sea clutter. The innovation of this paper is to propose a linear prediction error method for sea clutter to increase dynamic range of radar receiver. It prevents overloading and reaching saturation point at close range.

An efficient unsupervised band selection method based on an autocorrelation matrix for a hyperspectral image

In the field of unsupervised band selection, minimum linear prediction (LP) error is a commonly used criterion function. To avoid the large computational complexity, sequential forward selection (SFS) is often employed for subset search in LP-based methods. In this article, we propose a highly efficient LP-based band selection method termed autocorrelation matrix-based band selection (ACMBS), which adopts the sequential backward selection (SBS) as subset search strategy. Interestingly, the LP error is finally transformed into the inverse of the autocorrelation matrix in ACMBS. Thus the computational complexity of ACMBS is significantly reduced. Moreover, we further improve the accuracy of ACMBS by employing relative error, instead of absolute error, as a cost function which is invariant to the magnitude of bands. The results of the experiment show that ACMBS is quite efficient and outperforms the other compared methods in terms of classification accuracy as well.

Two-channel multi-stage speech enhancement for noisy fMRI environment

Strong acoustic noise interferes with speech communication during functional magnetic resonance imaging (fMRI) scans employing echo planar imaging (EPI) sequences. Performance of many speech enhancement (SE) methods degrades with such a low signal-to-noise ratio (SNR). In this paper, an adaptive SE method is proposed enabling noise-free online speech communication with patients and recording of clean speech. The proposed method is a two-channel, two-stage algorithm with a switching technique. In the first stage, least mean square (LMS) based algorithm and linear prediction error filtering, for specific parts of the noisy acoustic signal, are used. The first stage operations improve the performance of minimum mean square error log-spectral amplitude estimator (Log-MMSE) used for enhancing the speech signal in the second stage. A soft-switching technique is also employed which significantly reduces the audible noise spikes which occur during the transient periods between successive EPI sequencing. The proposed algorithm is tested online during the scans in a noisy fMRI room. Moreover, this paper provides the results of real time implementation of the algorithm on a dedicated laboratory test-bed setup using Texas Instruments digital signal processing platform.

Relationships of Measurement Error and Prediction Error in Observed‐Score Regression

The focus of this paper is assessing the impact of measurement errors on the prediction error of an observed‐score regression. Measures are presented and described for decomposing the linear regression's prediction error variance into parts attributable to the true score variance and the error variances of the dependent variable and the predictor variable(s). These measures are demonstrated for regression situations reflecting a range of true score correlations and reliabilities and using one and two predictors. Simulation results also are presented which show that the measures of prediction error variance and its parts are generally well estimated for the considered ranges of true score correlations and reliabilities and for homoscedastic and heteroscedastic data. The final discussion considers how the decomposition might be useful for addressing additional questions about regression functions’ prediction error variances.

On [formula omitted]-structure of bilinear models on [formula omitted

One-dimensionally indexed bilinear (BL) models are widely used for modeling non-Gaussian dataset. Extending BL models to multidimensionally indexed (spatial) (SBL) one yields a novel class of models which are capable of taking into account the non-Gaussianity character and spatiality behavior. Hence, the main contribution here is to study the L2-structure of some SBL models which play an important role in spatial statistical analysis. So, we establish necessary and sufficient conditions for the existence of regular second order stationary and ergodic solutions in terms of its transfer functions. As a consequence, we observe that the second order structure is similar to a weak ARMA field, and that the variance of the best linear prediction error is always greater than the one obtained from an SBL model.

Prediction error method for identification of LPV models

This paper is concerned with identification of linear parameter varying (LPV) systems in an input–output setting with Box–Jenkins (BJ) model structure. Classical linear time invariant prediction error method (PEM) is extended to the LPV PEM. Under the new LPV framework, identification of two types of input–output LPV models is considered: one is based on parameter interpolation and the other is based on model interpolation. The effectiveness of the proposed solution is validated by comparison with other existing LPV identification approaches through simulation examples and demonstrated by experiment studies.

Segmentation d’Images Texturées Couleur à l’aide de modèles paramétriques pour approche la distribution des erreurs de prédiction linéaires

We propose novel a priori parametric models to approximate the distribution of the two dimensional multichannel linear prediction error in order to improve the performance of color texture segmentation algorithms. Two dimensional linear prediction models are used to characterize the spatial structures in color images. The multivariate linear prediction error of these texture models is approximated with Wishart distribution and multivariate Gaussian mixture models. A novel color texture segmentation framework based on these models and a spatial regularization model of initial class label fields is presented. For the proposed method and with different color spaces, experimental results show better performances in terms of percentage segmentation error, in comparison with the use of a multivariate Gaussian law.

Capacity of OFDM Systems Over Fading Underwater Acoustic Channels

In this paper, we derive bounds to the channel capacity of orthogonal frequency division multiplexing (OFDM) systems over the underwater (UW) acoustic fading channel as a function of the distance between the transmitter and the receiver. The upper bound is obtained under perfect channel state information (CSI) at the receiver. The lower bound is obtained assuming the input is drawn from phase-shift keying (PSK) constellation which results in non-Gaussian distribution of the output signal and no CSI. The reduction from the upper bound is due to limited mutual information that can be conveyed by PSK constellation and the linear minimum mean square prediction error. Our UW channel deviates from the wide sense stationary and uncorrelated scattering (WSSUS) model commonly used for small bandwidths. We incorporate frequency-dependent path loss due to the acoustic propagation into each arrival path between the transmitter and the receiver. This leads the UW channel to be modeled as a frequency-dependent doubly spread fading channel characterized by the wide sense stationary and correlated scattering (WSS-non-US) fading assumption. Both Rayleigh and Ricean fading assumptions are investigated in our model. Results from the model show a gap between the upper and lower bounds which depends not only on the ranges and shape of the scattering function of the UW channel but also on the distance between the transmitter and the receiver. Our model for the scattering function was suggested by Rescheduled Acoustic Communications Experiment (RACE08) experimental data, leading to a multilag autoregressive (AR- q) model for the fading.

Method for reducing of noise by improving Signal-to-Noise-Ratio in wireless LAN

The signal to noise ratio (SNR) is one of the important measures for reducing the noise.A technique that uses a linear prediction error filter (LPEF) and an adaptive digital filter (ADF) to achieve noise reduction in a speech and image degraded by additive background noise is proposed. Since a speech signal can be represented as the stationary signal over a short interval of time, most of speech signal can be predicted by the LPEF. This estimation is performed by the ADF which is used as system identification. Noise reduction is achieved by subtracting the reconstructed noise from the speech degraded by additive background noise. Most of the MR image accelerating methods suffers from degradation of acquired images, which is often correlated with the degree of acceleration. However, Wideband MRI is a novel technique that transcends such flaws.In this paper we proposed LPEF and ADF for reducing the noise in speech and also we demonstrate that Wideband MRI is capable of obtaining images with identical quality as conventional MR images in terms of SNR in wireless LAN.

Parametric models of linear prediction error distribution for color texture and satellite image segmentation

In this article we present a Bayesian color texture segmentation framework based on the multichannel linear prediction error. Two-dimensional causal and non-causal real (in RGB color space) and complex (in IHLS and L ∗ a ∗ b ∗ color spaces) multichannel linear prediction models are used to characterize the spatial structures in color images. The main contribution of this segmentation methodology resides in the robust parametric approximations proposed for the multichannel linear prediction error distribution. These are composed of a unimodal approximation based on the Wishart distribution and a multimodal approximation based on the multivariate Gaussian mixture models. For the spatial regularization of the initial class label estimates, computed through the proposed parametric priors, we compare the conventional Potts model to a Potts model fusioned with a region size energy term. We provide performances of the method when using Iterated Conditional Modes algorithm and simulated annealing. Experimental results for the segmentation of synthetic color textures as well as high resolution QuickBird and IKONOS satellite images validate the application of this approach for highly textured images. Advantages of using these priors instead of classical Gaussian approximation and improved label field model are shown by these results. They also verify that the L ∗ a ∗ b ∗ color space exhibits better performance among the used color spaces, indicating its significance for the characterization of color textures through this approach.

A comparative study on chaoticity of equatorial/low latitude ionosphere over Indian subcontinent during geomagnetically quiet and disturbed periods

Abstract. In the present study, the latitudinal aspect of chaotic behaviour of ionosphere during quiet and storm periods are analyzed and compared by using GPS TEC time series measured at equatorial trough, crest and outside crest stations over Indian subcontinent, by employing the chaotic quantifiers like Lyapunov exponent (LE), correlation dimension (CD), entropy and nonlinear prediction error (NPE). It is observed that the values of LE are low for storm periods compared to those of quiet periods for all the stations considered here. The lowest value of LE is observed at the trough station, Agatti (2.38° N, Geomagnetically), and highest at crest station, Mumbai (10.09° N, Geomagnetically) for both quiet and storm periods. The values of correlation dimension computed for TEC time series are in the range 2.23–2.74 for quiet period, which indicate that equatorial ionosphere may be described with three variables during quiet period. But the crest station Mumbai shows a higher value of CD (3.373) during storm time, which asserts that four variables are necessary to describe the system during storm period. The values of non linear prediction error (NPE) are lower for Agatti (2.38° N, Geomagnetically) and Jodhpur (18.3° N, Geomagnetically), during storm period, compared to those of quiet period, mainly because of the predominance of non linear aspects during storm periods The surrogate data test is carried out and on the basis of the significance of difference of the original data and surrogates for various aspects, the surrogate data test rejects the null hypothesis that the time series of TEC during storm and quiet times represent a linear stochastic process. It is also observed that using state space model, detrended TEC can be predicted, which reasonably reproduces the observed data. Based on the values of the above quantifiers, the features of chaotic behaviour of equatorial trough crest and outside the crest regions of ionosphere during geomagnetically quiet and disturbed periods are briefly discussed.

Epileptic EEG detection using the linear prediction error energy

In this study, a method is proposed to detect epileptic seizures over EEG signal. For this purpose, a linear prediction filter is used to observe the presence of spikes and sharp waves on seizure EEG recordings. Linear prediction analysis calculates a coefficient set for each window, which can best model the applied time series signal. Modeling success is observed on the prediction error signal. The presence of spikes and other seizure-specific sharp waves on the signal reduces the modeling success and increases the prediction error of the filter. It is clearly observed that, the energy of prediction error signal during seizures is much higher than that of the seizure free intervals, which indicates the energy value and can be used to locate the seizure interval. The method is applied to 250 distinct EEG records, each of which has 23.6 s duration. The results of the proposed algorithm are evaluated with the ROC analysis which indicates 93.6% success in detecting the presence of seizures. As a conclusion, the linear prediction error energy method can be considered as an efficient way to detect epileptic seizures on EEG records.

낮은 계산 복잡도를 갖는 Linear Prediction 기반의 SNR 추정 기법

It is very important to estimate the Signal to Noise Ratio(SNR) of received signal in time varying channel state. Most SNR estimation techniques derive the SNR estimates solely from the samples of the received signal after the matched filter. In the severe distorted wireless channel, the performance of these estimators become unstable and degraded. LP-based SNR estimator which can operate on data samples collected at the front-end of a receiver shows more stable performance than other SNR estimator. In this paper, we study an efficient SNR estimation algorithm based on LP and propose a new estimation method to decrease the computation complexity. Proposed algorithm accomplishes the SNR estimation process efficiently because it uses the forward prediction error and its conjugate value during the linear prediction error update. Via the computer simulation, the performance of this proposed estimation method is compared and discussed with other conventional SNR estimators in digital communication channels.

Linear prediction based semiblind channel estimation for multiuser OFDM with insufficient guard interval

To meet the demand of high data rate transmissions for multimedia wireless communications, orthogonal frequency division multiplexing (OFDM) systems in conjunction with multiple-input multiple-output (MIMO) signal processing have been considered one of the central techniques in advanced wireless communications. In the paper, two semiblind channel estimation algorithms are proposed for the uplink multiuser OFDM systems with insufficient guard interval, in contrast to sufficient guard interval assumed in most of the prior works. A zero-padding OFDM system, which zero-pads rather than cyclicly prefixing each block, is considered in this paper. By utilizing the relation between the linear prediction error filters (LPEFs) of the received signal with multiple prediction orders and the transmitted data sequence, the first proposed algorithm, namely the multistage LP (MLP) based algorithm, can estimate the MIMO channel coefficients, with only a single pilot OFDM block used. To reduce the sensitivity of the proposed algorithms to the channel order overestimation, it is proposed to implement the LPEFs with a QR-decomposition based approach. This QRdecomposition based approach alternatively computes the LPEFs without direct inversion of the received signal correlation matrix, thus exhibiting robustness against channel order overestimation. Some simulation results are presented to demonstrate the effectiveness and robustness of the proposed algorithms.

Modeling and System Identification of the muFly Micro Helicopter

An accurate mathematical model is indispensable for simulation and control of a micro helicopter. The nonlinear model in this work is based on the rigid body motion where all external forces and moments as well as the dynamics of the different hardware elements are discussed and derived in detail. The important model parameters are estimated, measured or identified in an identification process. While most parameters are identified from test bench measurements, the remaining ones are identified on subsystems using the linear prediction error method on real flight data. The good results allow to use the systems for the attitude and altitude controller design.