This letter provides a solution to a robust adaptive beamformer used in remote sensing systems under the effect of two factors: i) the impulsive noise satisfying Gaussian mixture distribution and ii) unknown model mismatches, such as look direction errors, sensor location perturbations, etc. First, a beamformer output-based cost function is designed to effectively suppress the influence of impulsive noise. Second, the weight vector of the beamformer is regularized through norm constraints to reduce the impact of model mismatches. Furthermore, an efficient iterative scheme is established to quickly minimize the designed cost function based on their properties, and a robust adaptive beamformer can be obtained. Finally, computer simulations are presented to demonstrate the robust performance and fast convergence of the novel algorithm in the presence of impulsive noise and model mismatches.

Abstract The goal of feature matching is to establish accurate correspondences between feature points in different images depicting the same scene. To address the polymorphism of local structures, the authors propose a mismatch removal method using bilateral local–global structural consistency. This method incorporates the problem of mismatch removal into the framework of graph matching, constructs a global affinity matrix using local structural similarity and global affine transformation consistency, and optimizes it using a constrained integer quadratic programming method. To comprehensively describe the local structure, the signature quadratic form distance (SQFD) is used to measure the consistency of the neighbourhood structure. Specifically, the weights of edges are constructed based on the SQFD of the local structure, while the matching correctness of nodes and edges between the two graphs is described using local vector similarity. Furthermore, the consistency of the global affine transformation is evaluated by assessing the consistency of the local neighbourhood affine transformation between different corresponding point pairs. In estimating the local affine transformation, a bilateral correction is performed using a total least‐squares (TLS) algorithm to measure the similarity of nodes between the two different graphs. Experimental results demonstrate that the proposed algorithm outperforms state‐of‐the‐art methods in terms of accuracy and effectiveness.

Numerous methods have exploited projection learning to extract low-dimensional features for image classification. Some projection learning methods integrate low-rank matrix recovery into classification models to equip the projection subspace with discrimination and robustness against corruption. However, these methods cannot directly recover “clean” components from the new datum in a low-dimensional subspace. Additionally, they are sensitive to the selection of projection dimensions. To overcome these shortcomings, we propose a dual discriminative low-rank projection learning framework for robust image classification. Specifically, the proposed method learns a low-rank projection and a semi-orthogonal projection to recover “clean” components from the original data and simultaneously obtain a low-dimensional subspace. Thereafter, to preserve discriminative information in the low-dimensional subspace, an$L_{2,1}$-norm term is constructed by concentrating the projected intra-class samples around their adaptive class centroids. Regression-based terms are appended using the low-dimensional features extracted from the recovered clean data and the class centroids for more accurate classification. Experiments on five public databases with various corruptions demonstrate that the proposed method can robustly classify image data despite a small training sample sizes and gross corruption. The superiority of the proposed method is further verified on the large-scale PubFig83 database, on which it achieves an 87.58% classification accuracy.

In wireless communication systems, the conventional constant modulus algorithm incurs artificial error and steady state misadjustment. In this study, an improved constant modulus algorithm (ICMA) and its generalized form are proposed for blind equalization. The ICMA utilizes the clustering function of Gaussian function to efficiently suppress this artificial error and steady state misadjustment at the cost of reduction in the sample usage rate. Moreover, the generalized form of the ICMA is developed to ensure the sample usage rate while the good performances of the ICMA are maintained. Simulation results illustrate better equalization performances of the ICMA and generalized ICMA, as compared to the classical constant modulus algorithm.

Edge detection is a critical component of synthetic aperture radar (SAR) image interpretation. Due to serious speckle noise, the core problems for SAR edge detection are how to keep a constant false alarm rate (CFAR) and how to achieve unbiased localization of the edges. Aiming at these problems, this article proposes a novel edge detector with a unique structure for noise-contaminated SAR images, which creatively integrates the difference operation with ratio operation (hence named as “UDR: unbiased difference-ratio” edge detector). Theoretical analysis proves that the difference operation effectively affords the UDR unbiased localization ability for both ideal and nonideal edges, and the ratio operation provides the UDR the property of CFAR under the influence of speckle noise. Experimental results on both simulated and real-world SAR images demonstrate the unbiased localization ability of the proposed UDR edge detection, insensitive to the changes of edge contrast, the width of the transition zone and the noise level. Benefited from the superior localization precision and insensitivity to noise, the average true positive detection rate of the proposed detector is improved to 95%, outperforming the compared state-of-the-art methods.

High-dimensional small sample size problems exist in the real world, which significantly increases the difficulty of data processing. In this paper, we propose a kernel-based within class collaborative preserving discriminant projection method to reduce data dimensionality. In order to deal with nonlinear problems and improve the discrimination of the projection subspace, the proposed method preserves the collaborative reconstruction relationship of the same class samples in the kernel space, and pursuits maximizing the between class scatter. A two-step eigenvalue decomposition method is used to stably obtain the optimal discriminant projection matrix. Moreover, simulation experiments show that, even in low dimensions and small sample size, the proposed method can achieve high recognition accuracy.

In this letter, an antistretch edge detector has been proposed for synthetic aperture radar (SAR) images. Traditional detectors using anisotropy edge detection filters often incur severe edge stretch. If isotropy edge detection filters are used, the detectors usually have poor edge resolution. Hence, we skillfully fuse an anisotropy edge detection filter with an isotropy one. By embedding the fused filter into the routine ratio-based SAR edge detector, an antistretch edge detector is proposed. Benefiting from the fused edge detection filter, the proposed edge detector can obtain a good antistretch ability and keep a high edge resolution. A theoretical analysis indicates that the computational complexity of the proposed edge detector is close to conventional ratio-based edge detectors. The fusion operation will not affect the constant false alarm rate property. The receiver-operating-characteristic curves are used to objectively evaluate the proposed detector. Experimental studies show that the proposed detector has a lower false positive rate than the majority of detectors using only anisotropy filters or isotropy filters. Furthermore, the experimental results on simulated and real-world SAR images show that the proposed antistretch edge detector can obtain an accurate edge map.

Multiple-input multiple-output radar has many advantages over the phased-array radar system due to the waveform diversity, one of which is the greater flexibility to design the transmit beampattern. Hence, waveform set design for transmit beampattern has become an attractive topic, and many methods have been proposed in recent years. However, previous methods cannot synthesize the waveforms with constant-envelope and easy-generation properties. In this letter, we propose to design a set of correlated linear frequency modulation (LFM) waveforms to solve this problem. First, the covariance matrix of the LFM waveform set is analyzed, and the formulation of the transmit beampattern is obtained accordingly. Since the transmit beampattern is mainly affected by the frequency steps and initial phases of the LFM waveforms, the correlated LFM waveform set design problems are formulated by optimizing these parameters for different beampatterns. The resulting problems are solved by adopting the constrained nonlinear optimization, and the LFM waveforms are obtained consequently. The designed waveforms have the properties of constant-envelope and easy generation, and can match the desired transmit beampattern properly. Simulation results demonstrate the superiority of our proposed method.

The orthogonal frequency-division multiplexing (OFDM) chirp waveform has attracted much attention due to its high range resolution, low peak-to-average ratio, and large time-bandwidth product. In its application to multiple-input multiple-output radar, the correlation property of multiple OFDM chirp waveforms should be considered primarily for good detection performance. The simulation results show that high sidelobes exist in correlation functions of the conventional waveforms. In this letter, the reason for high correlation sidelobes is explored first, which is the equal subchirp durations and the same subcarrier bandwidth. Second, two new OFDM chirp waveform diversity design schemes are proposed to suppress the correlation interference. Via designing the various subchirp durations or subcarrier bandwidths specially, the high sidelobes are depressed and the correlation property is improved. Both simulation results and comparisons verify the effectiveness of the proposed methods.

This letter proposes a rapid line-extraction (RLE) method for synthetic aperture radar (SAR) images. RLE first transforms an image in the space domain into an image in the frequency domain. Then, using the central-slice theorem, RLE skilfully maps the image in the frequency domain into a parameter space, which effectively accelerates the straight-line extraction process. Unlike the traditional Hough transform, RLE is performed directly on an edge-field image rather than on a binary edge map. Theoretical analysis proves the advantages of using the edge-field map. Notably, the computational complexity can be greatly reduced relative to the complexity of obtaining a binary edge map, and the method can efficiently avoid the negative influence of false edges in the binary edge map. More importantly, because speckle, clutter, and blurred edges in real-world images decrease the sharpness of peaks, edge-field images that include the strength and direction information of SAR images are adopted to reduce the diffusion of peaks and improve the detection accuracy. Experimental studies show that RLE works independently, is robust to noise, has low computational complexity, achieves high true-positive detection rates, and yields satisfactory detection precision.

This paper introduces an efficient soft decision-directed algorithm (SDDA) for blind equalization of 4 quadrature amplitude modulation (4-QAM) systems. A novel cost function (CF) based on the conventional SDDA is established to efficiently obtain the weight vector of the blind equalizer (BE). An appropriately modified Newton method (MNM) which is proposed in our previous work [9] is adopted to fast search the optimal equalizer weight vector. It is shown that the BE based on the novel MNM has an approximately quadratic order of convergence like Newton methods. Moreover, the proposed algorithm has better stability and much lower computational load than Newton methods.

BACKGROUND: The planted (l, d) motif search (PMS) is an important yet challenging problem in computational biology. Pattern-driven PMS algorithms usually use k out of t input sequences as reference sequences to generate candidate motifs, and they can find all the (l, d) motifs in the input sequences. However, most of them simply take the first k sequences in the input as reference sequences without elaborate selection processes, and thus they may exhibit sharp fluctuations in running time, especially for large alphabets. RESULTS: In this paper, we build the reference sequence selection problem and propose a method named RefSelect to quickly solve it by evaluating the number of candidate motifs for the reference sequences. RefSelect can bring a practical time improvement of the state-of-the-art pattern-driven PMS algorithms. Experimental results show that RefSelect (1) makes the tested algorithms solve the PMS problem steadily in an efficient way, (2) particularly, makes them achieve a speedup of up to about 100× on the protein data, and (3) is also suitable for large data sets which contain hundreds or more sequences. CONCLUSIONS: The proposed algorithm RefSelect can be used to solve the problem that many pattern-driven PMS algorithms present execution time instability. RefSelect requires a small amount of storage space and is capable of selecting reference sequences efficiently and effectively. Also, the parallel version of RefSelect is provided for handling large data sets.

In this study, a robust adaptive beamformer against large direction‐of‐arrival (DOA) mismatch for multiple‐input–multiple‐output radar is proposed with linear phase and magnitude constraints on main lobe. First, the full‐dimensional weight vector (WV) is expressed as the Kronecker product of the transmit and receive array WVs based on the WV separable principle. For the transmit array WV, the authors find an interesting property that the Fourier spectrum of its conjugate inverse arrangement is equal to its array response function within a phase factor. This property also exists in the receive array WV. Using this property, the phase response of the transmit and receive array, respectively, is set to be linear based on designing a finite impulse response filter. Then, a bi‐quadratic cost function with respect to the transmit and receive WVs is established by only constraining the real magnitude response and it is effectively solved by the bi‐iterative algorithm. The proposed beamformer has lower computational complexity and faster sample convergence rate, compared with the traditional magnitude response constraints beamformers with full degrees of freedom. Moreover, it can provide good robustness against large DOA mismatch. Numerical experiments are provided to demonstrate the effectiveness of the proposal.

Conventional line detection methods are mainly based on the binary edge map. This letter proposes a new line detection method that directly extracts line features from the image edge fields of the synthetic aperture radar (SAR) images. In the proposed method, the strength and direction of each field point are first obtained using a ratio-based edge filter. Then, the accumulation weight of the field point is jointly computed using its strength and direction. The direction of a field point on the line is essentially the orientation of the line. Furthermore, a field point on a strong line should be distinguished from a field point on a weak line. Thus, the accumulation weights of different field points are not equal. By summing up the accumulation weights, the straight lines in the SAR image space are directly converted into several local peaks in the parameter space. A sort-window peak detection method is proposed to suppress the spurious secondary peaks in the parameter space. The experimental results show that the proposed line detection method is robust to noise and has a good antiocclusion ability. The proposed method performs well in terms of true positive detection rate and detection accuracy for both synthetic and real-world images.

By introducing a crater-shaped window (CSW) instead of the traditional square-shaped window (SSW), an edge detector with low false positive rate is proposed to rapidly extract thin edges of synthetic aperture radar images. For further refining the true positive rate, we skillfully introduce an edge compensation strategy. Using the CSW, the square successive difference of averages is calculated. Then, edge compensation strategy is used. The CSW has low sidelobe and high localization accuracy to ensure that a detector has low false positive rate. Moreover, by using edge compensation strategy, boundaries between two similar homogeneous areas can be easily extracted. The proposed detector, hence, has high true positive rate. Both objective and subjective experiment results show that the edge detector using CSW and having edge compensation strategy attains better performance than one using SSW and without edge compensation strategy.

The planted (l, d) motif search (PMS) is an important yet challenging problem in computational biology. Patterndriven PMS algorithms usually use k out of t input sequences as reference sequences to generate candidate motifs, and they can find all the (l, d) motifs in the input sequences. However, most of them simply take the first k sequences in the input as reference sequences without elaborate selection processes, and thus they may exhibit sharp fluctuations in running time, especially for large alphabets. In this paper, we build the reference sequence selection problem and propose a method named RefSelect to quickly solve it by evaluating the number of candidate motifs for the reference sequences. RefSelect can bring a practical time improvement of the state-of-the-art pattern-driven PMS algorithms. Experimental results show that RefSelect (1) makes the tested algorithms solve the PMS problem steadily in an efficient way, (2) particularly, makes them achieve a speedup of up to about 100× on the protein data, and (3) is also suitable for large data sets which contain hundreds or more sequences.

This paper addresses the problem of adaptive source extraction via the canonical correlation analysis (CCA) approach. Based on Liu's analysis of CCA approach, we propose a new criterion for source extraction, which is proved to be equivalent to the CCA criterion. Then, a fast and efficient online algorithm using quasi-Newton iteration is developed. The stability of the algorithm is also analyzed using Lyapunov's method, which shows that the proposed algorithm asymptotically converges to the global minimum of the criterion. Simulation results are presented to prove our theoretical analysis and demonstrate the merits of the proposed algorithm in terms of convergence speed and successful rate for source extraction.

This paper proposes a space-time semi-blind equalizer (ST-SBE) for dispersive multiple-input multiple-output (MIMO) communication systems that employ high throughput quadrature amplitude modulation (QAM) signals. A novel cost function (CF) that integrates multimodulus algorithm (MMA) with soft decision-directed (SDD) scheme is established to efficiently obtain the weight vector associated with the ST-SBE. In the ST-SBE, a very short training sequence is used to provide a rough initial least squares estimate of the weight vector. An efficient modified Newton method (MNM) for minimizing the established cost function is proposed to fast search the optimal weight vector. Very interestingly, we prove that the proposed MNM has the same quadratic order of convergence as Newton methods. In addition, the proposed MNM has much lower computational complexity than Newton methods. Simulation results are provided to demonstrate that the ST-SBE has better performances than the gradient-Newton (GN)-based concurrent constant modulus algorithm (CMA) with SDD scheme (GN-CMA+SDD).

The problem of blind separation of complex-valued signals via joint diagonalization of a set of non-unitary target matrices is addressed in this paper. An improved blind source separation (BSS) algorithm is developed based on minimization of the Frobenius-norm formulation of the approximate joint diagonalization problem by using a multiplicative update. Such minimization yields a strictly diagonally-dominant updated matrix at each iteration. With relaxing some constraints on the target matrices, the improved BSS algorithm allows for extended applications. The behavior of the improved BSS algorithm is demonstrated by computer simulation results in comparison with some representative BSS algorithms.

This paper is concerned with blind separation of convolutive sources. The main idea is to make an explicit exploitation of block Toeplitz structure and block-inner diagonal structure in autocorrelation matrices of source signals at different time delays as well as of inherent relations among these matrices. With implementation of joint block diagonalization, a tri-quadratic cost function is introduced so that the mixture matrix can be extracted from a set of the correlation matrices of the observed vector sequence without pre-whitening. In this novel one-stage algorithm, every iteration step involves finding the closed solution to the corresponding least squares problem. Once the estimate of the mixing matrix is obtained, the source signals are retrieved by the classical least squares methods. The performance of the proposed algorithm is illustrated by simulation results.

We propose an adaptive improved natural gradient algorithm for blind separation of independent sources. First, inspired by the well-known backpropagation algorithm, we incorporate a momentum term into the natural gradient learning process to accelerate the convergence rate and improve the stability. Then an estimation function for the adaptation of the separation model is obtained to adaptively control a step-size parameter and a momentum factor. The proposed natural gradient algorithm with variable step-size parameter and variable momentum factor is therefore particularly well suited to blind source separation in a time-varying environment, such as an abruptly changing mixing matrix or signal power. The expected improvement in the convergence speed, stability, and tracking ability of the proposed algorithm is demonstrated by extensive simulation results in both time-invariant and time-varying environments. The ability of the proposed algorithm to separate extremely weak or badly scaled sources is also verified. In addition, simulation results show that the proposed algorithm is suitable for separating mixtures of many sources (e.g., the number of sources is 10) in the complete case.

The problem of blind source separation (BSS) using joint diagonalization of a set of non-unitary eigen-matrices that are obtained with the observed signal vector sequence is addressed in this paper. A theoretical study is conducted of the identifiability of joint diagonalization of non-orthogonal matrices so as to generalize some known results for the orthogonal case. In particular, a mathematical proof is provided for essential uniqueness of general joint diagonalization, that is to say, all the estimated mixing matrices extracted from the non-unitary eigen-matrix group are essentially equal within an arbitrary permutation and scaling. The non-orthogonal identifiability theorem given in this paper serves as a mathematical foundation for the BSS methods based on the non-orthogonal joint diagonalization.

We propose an analytic search method for interferometric synthetic aperture radar (InSAR) image registration. An analytic cost function is established, and the subpixel offsets associated with the maximum of the cost function are continuously searched by direction-alternate optimization methods along vertical and horizontal directions. We establish a novel dual-quartic analytic cost function for extraction of subpixel offsets from a corresponding pixel pair established by pixel-level registration. A robust optimization method that is called biiteration algorithm for searching the maximum point of the dual-quartic cost function is developed to solve the subpixel offsets. Since the dual-quartic cost function is quartic if one of two parameters is fixed, one of two substeps including in each iterative step of the efficient biiterative method is to find the solution by the secant method. The performances of the proposed method are shown by using simulated and real data and compared with those of representative existing registration method.

Adaptive IIR filtering in the case where noise exists in both the input and output of the system amounts to solving over-determined normal equations. In this paper a recursive total instrumental-variable (RTIV) algorithm is proposed for tracking the total least-squares (TLS) solution of the normal equations in the over-determined instrumental-variable methods. It is shown that the weight vector in the RTIV algorithm converges to the direction parallel to the singular vector associated with the smallest singular value of the augmented cross-correlation matrix. Moreover, the estimated parameters of the adaptive IIR filter are optimal in the TLS sense and its noise rejection capability is superior to that of the least-squares based algorithms. The appealing behavior of the RTIV algorithm for noisy adaptive IIR filtering is substantiated by simulation results.

This paper is concerned with FIR filtering with noise-corrupted input-output measurements. With an analysis of the algebraic structure of the correlation matrix, it is shown that an unbiased estimate of FIR parameters can be obtained by solving a special bilinear equation. Then a bilinear equation method (BEM) is developed for solving the bilinear equation associated with the unbiased solution of the FIR filtering under the unknown ratio of the input noise variance to the output noise variance (NNR). Being different from the existing unbiased estimators, the main advantage is that the proposed method exploits much sufficiently the special structure of the correlation matrix and obtains much accurate estimation for FIR filtering in the presence of input and output noises. Simulation results are presented to validate the good performance of the proposed method.

In this paper, we develop a fast recursive algorithm with a view to finding the total least squares (TLS) solution for adaptive FIR filtering with input and output noises. We introduce an approximate inverse power iteration in combination with Galerkin method so that the TLS solution can be updated adaptively at a lower computational cost. We further reduce the computational complexity of the developed algorithm by making efficient computation of the fast gain vector. We then make a careful investigation into global convergence of the developed algorithm. Simulation results are provided that clearly illustrate appealing performances of the developed algorithm.

In this paper an improved whitening scheme is first developed by estimating the signal subspace jointly from a set of diagonalization-structural matrices based on the proposed cyclic maximizer of an interesting cost function. Next, a biquadratic contrast function is proposed for extracting one single independent component from a slice matrix group of any order cumulant of the array signals in the presence of the spatially-temporally white noise. A fast fixed-point algorithm is constructed for searching a minimum point of the proposed contrast function. Then multiple independent components are obtained by using repeatedly the fixed point algorithm for extracting one single independent component, and the orthogonality among them is achieved by the well-known QR decomposition. The performance of the proposed algorithms is illustrated by simulation results and is compared with several representative blind source separation algorithms.

We propose an adaptive filtering approach for interferograms, which is a modification to the Lee adaptive complex filter. Based on local frequency estimates, we compute the normal orientation of local phase fringes. A directionally dependent filtering window is aligned perpendicular to the normal orientation of local phase fringes (i.e., along local phase fringes) by interpolation, making the pixels included in the filtering window have approximately more homogeneous values. Moreover, the computation of the filter parameter does not require local phase unwrapping in the real plane. This filter minimizes the loss of signal and reduces the level of noise. By using two sets of simulated data, its effectiveness can be seen in terms of the fidelity to noise-free phases, fringe preservation, and residue reduction.

A novel random-gradient-based algorithm is developed for online tracking the minor component (MC) associated with the smallest eigenvalue of the autocorrelation matrix of the input vector sequence. The five available learning algorithms for tracking one MC are extended to those for tracking multiple MCs or the minor subspace (MS). In order to overcome the dynamical divergence properties of some available random-gradient-based algorithms, we propose a modification of the Oja-type algorithms, called OJAm, which can work satisfactorily. The averaging differential equation and the energy function associated with the OJAm are given. It is shown that the averaging differential equation will globally asymptotically converge to an invariance set. The corresponding energy or Lyapunov functions exhibit a unique global minimum attained if and only if its state matrices span the MS of the autocorrelation matrix of a vector data stream. The other stationary points are saddle (unstable) points. The globally convergence of OJAm is also studied. The OJAm provides an efficient online learning for tracking the MS. It can track an orthonormal basis of the MS while the other five available algorithms cannot track any orthonormal basis of the MS. The performances of the relative algorithms are shown via computer simulations.

An adaptive algorithm for real-time channel estimation is proposed, which is a blind estimation of signal-to-interference ratio and signal-to-noise ratio. It works well in the case of multiple interferences and noise by using an improved adaptive k-NN technique. The channel in wireless security communications, such as adaptive frequency hopping (AFH), would be taken as an example to show the validity of the proposed algorithm. The algorithm is particularly effective in the case of noise-free signal disturbed by interferences of the known type. In our discussion, the signal is /spl pi//4 DQPSK, and the interferences and noise are GFSK, CW, and AWGN. The reason why we assume the interferences and noise are GFSK, CW and AWGN is discussed in the full paper. When the signal or interferences are of the other types, the estimator given in the full paper also brings good results. The algorithm has been implemented on the platform based on ADSP21160, and it works well.

This paper proposes a novel cross-correlation neural network (CNN) model for finding the principal singular subspace of a cross-correlation matrix between two high-dimensional data streams. We introduce a novel nonquadratic criterion (NQC) for searching the optimum weights of two linear neural networks (LNN). The NQC exhibits a single global minimum attained if and only if the weight matrices of the left and right neural networks span the left and right principal singular subspace of a cross-correlation matrix, respectively. The other stationary points of the NQC are (unstable) saddle points. We develop an adaptive algorithm based on the NQC for tracking the principal singular subspace of a cross-correlation matrix between two high-dimensional vector sequences. The NQC algorithm provides a fast online learning of the optimum weights for two LNN. The global asymptotic stability of the NQC algorithm is analyzed. The NQC algorithm has several key advantages such as faster convergence, which is illustrated through simulations.

This paper proposes a cross-associative neural network (CANN) for singular value decomposition (SVD) of a non-squared data matrix in signal processing, in order to improve the convergence speed and avoid the potential instability of the deterministic networks associated with the cross-correlation neural-network models. We study the global asymptotic stability of the network for tracking all the singular components, and show that the selection of its learning rate in the iterative algorithm is independent of the singular value distribution of a non-squared matrix. The performances of CANN are shown via simulations.