Search

Hyperspectral images (HSIs) provide rich spectral information, facilitating many applications, including landcover classification. However, due to the high dimensionality of HSIs, landcover mapping applications usually suffer from the curse of dimensionality, which degrades the efficiency of supervised classifiers due to insufficient training samples. Feature extraction (FE) is a popular dimension reduction strategy for this issue. This paper proposes an unsupervised FE algorithm that involves extracting endmembers and clustering spectral bands. The proposed method first extracts existing endmembers from the HSI data via a vertex component analysis method. Using these endmembers, it subsequently constructs a prototype space (PS) in which each spectral band is represented by a point. Similar/correlated bands in the PS remain near one another, forming several clusters. Therefore, our method, in the next step, clusters spectral bands into multiple clusters via K-means and fuzzy C-means algorithms. Finally, it combines all the spectral bands in the same cluster using a weighted average operator to decrease the high dimensionality. The extracted features were evaluated by applying an SVM classifier. The experimental results confirmed the superior performance of the proposed method compared with five state-of-the-art dimension reduction algorithms. It outperformed these algorithms in terms of classification accuracy on three widely used hyperspectral images (Indian Pines, KSC, and Pavia Centre). The suggested technique also showed comparable or even stronger performance (up to 9% improvement) compared with its supervised competitor. Notably, the proposed method exhibited higher accuracy even when only a limited number of training samples were available for supervised classification. Using only five training samples per class for the KSC and Pavia Centre datasets, our method’s classification accuracy was higher than that of its best-performing unsupervised competitors by about 7% and 1%, respectively, in our experiments.

In this paper, we develop a new algorithm for centralized target detection in passive MIMO radar (PMR) using sparse recovery technique. PMRs use a network of receivers and illuminators of opportunity to detect and localize targets. We consider a widely separated PMR network assuming the availability of reference channels. We first transform the collected information of all receivers to a common space and combine them to attain a unified model. The problem of target detection in the extracted model is equal to a block sparse recovery problem. Since employing the generic sparse recovery tools are impractical due to the ultra-large dimension of the sensing matrix, we exploit the structure of the involving matrices and propose a very efficient distributed algorithm which extracts all scatterers, including targets and clutter simultaneously with a unified procedure. The proposed algorithm is highly efficient, and it does not require a high bandwidth link to transfer raw data from nodes to the fusion center. Moreover, the algorithm inherently benefits from parallel processing and distributes the extensive computations among receivers. Our simulation results demonstrate that the proposed algorithm outperforms the popular PMR detection algorithm, especially in the presence of interfering targets and any strong clutter residue.

Abstract The distribution of the received signals in many array processing applications is noncircular. Although optimal widely linear beamformer (WLB) can provide the best performance for noncircular received signals, its performance degrades severely under model mismatches in practical applications. As a remedy, we propose a robust WLB by using precise reconstruction of extended interference-plus-noise covariance matrix (EINCM) and low-complexity estimation of extended desired signal steering vector (EDSSV). We propose to first determine the steering vectors, powers, and noncircularity coefficients of all signals and the noise power. In contrast to the previous reconstruction methods using the integration over a wide angular sector, we reconstruct the interference-plus-noise covariance matrix (INCM) and the pseudo INCM accurately according to their definitions. By using INCM and pseudo INCM, we can precisely reconstruct the EINCM. We propose to estimate the EDSSV by intersecting two extended subspaces, which are respectively formed by eigendecomposing the extended sample covariance matrix and the extended desired signal covariance matrix. Unlike the convex optimization methods, the proposed EDSSV estimation does not require any optimization programming and yields a solution with closed expression in low computational complexity. Simulation results show that the proposed robust WLB provides near optimal performance under several model mismatch cases.

Cardiac mapping systems are based on the time/frequency feature analyses of intracardiac electrograms recorded from individual bipolar/unipolar electrodes. Signals from each electrode are processed independently. Such approaches fail to investigate the interrelationship between simultaneously recorded channels of any given mapping catheter during atrial fibrillation (AF). We introduce a novel signal processing technique that reflects regional dominant frequency (RDF) components. We show that RDF can be used to identify and characterize variation and disorganization in wavefront propagation- wave breaks. The intracardiac electrograms from the left atrium of 15 patients were exported to MATLAB and custom software employed to estimate RDF and wave break rate (WBR). We observed a heterogeneous distribution of both RDF and WBR; the two measures were weakly correlated (0.3; p < 0.001). We identified locations of AF or atrial tachycardia (ATach) termination and later compared offline with RDF and WBR maps. We inspected our novel metrics for associations with AF termination sites. Areas associated with AF termination demonstrated high RDF and low WBR (↑RDF,↓WBR). These sites were present in 14 of 15 patients (mean 2.6 ± 1.2 sites per patient; range, 1–4 sites), 43% situated within the pulmonary veins. In nine patients where AF terminated to sinus rhythm (6) or ATach (3), post-hoc analysis demonstrated all ↑RDF,↓WBR sites were ablated and correlated with AF termination sites. The proposed RDF signal processing tools can be used to identify and quantify wave break, and the combined use of these two novel metrics can aid characterization of AF. Further prospective studies are warranted.

This paper presents a novel noise-robust feature extraction method for speech recognition using the robust perceptual minimum variance distortionless response (MVDR) spectrum of temporally filtered autocorrelation sequence. The perceptual MVDR spectrum of the filtered short-time autocorrelation sequence can reduce the effects of residue of the nonstationary additive noise which remains after filtering the autocorrelation. To achieve a more robust front-end, we also modify the robust distortionless constraint of the MVDR spectral estimation method via revised weighting of the subband power spectrum values based on the sub-band signal to noise ratios (SNRs), which adjusts it to the new proposed approach. This new function allows the components of the input signal at the frequencies least affected by noise to pass with larger weights and attenuates more effectively the noisy and undesired components. This modification results in reduction of the noise residuals of the estimated spectrum from the filtered autocorrelation sequence, thereby leading to a more robust algorithm. Our proposed method, when evaluated on Aurora 2 task for recognition purposes, outperformed all Mel frequency cepstral coefficients (MFCC) as the baseline, relative autocorrelation sequence MFCC (RAS-MFCC), and the MVDR-based features in several different noisy conditions.

Channel models for outdoor wireless systems usually assume two-dimensional (2D) random scattering media. In the practical outdoor wireless channels, the impact of the wave propagation in the third-dimension is definitely important; especially when the communication system efficiently exploits potentials of multiple antennas. In this paper, we propose a new model for multiple-input multiple-output (MIMO) multicarrier propagation channels in a three-dimensional (3D) environment. Specifically, the proposed model describes the cross-correlation function (CCF) between two subchannels of an outdoor MIMO channel employing directional antennas and in the presence of nonisotropic wave propagation in 3D space. The derived CCF consists of some correlation terms. Each correlation term is in the form of a linear series expansion of averaged Bessel functions of the first kind with different orders. In practice, each correlation term has a limited number of Bessel components. Our numerical evaluations show the impact of different parameters of the propagation environment as well as the employed antennas on the resulting CCF. Using the proposed CCF, we also establish simple formulas to approximate the coherence time, the coherence bandwidth and the spatial coherence of such channels. The numerical curve fitting results fit to the empirical results reported in the channel modeling literature.