Sparse Representation Framework Research Articles

Mixed sparse representation for approximated observation-based compressed sensing radar imaging

Recently, compressed sensing (CS) has been applied in synthetic aperture radar (SAR). A framework of mixed sparse representation (MSR) has been proposed for reconstructing SAR images due to the complicated ground features. The existing method decomposes the image into the point and smooth components, where the sparse constraint is directly applied to the smooth components. This makes it difficult to tackle the complex-valued SAR images, since the phase angles of SAR images are always stochastic. A magnitude-phase separation MSR method is proposed for CS-SAR imaging based on approximated observation. Compared to the existing method, the proposed method has better reconstruction ability, because it only imposes the sparse constraint on the magnitude of the smooth components, and therefore, the phase angles are still stochastic. Furthermore, owing to the inherent low memory requirement of approximated observation, the proposed method requires much less memory cost. In the simulation and experimental results, the proposed method deals with the complex-valued SAR images effectively and demonstrates superior performance over the chirp scaling algorithm and the existing MSR method.

Compressed Sensing of Underwater Acoustic Signals via Structured Approximation $l_0$-Norm

The underwater wireless sensor networks enable the telemonitoring and communications technologies of underwater information play important roles in the gathering process of scientific data in collaborative monitoring missions. However, to design such a system, several aspects must be considered to make fewer resources required, such as energy efficiency, miniaturization, the required functionality, etc. Conventional methods of data compression and reconstruction fail in energy efficiency. Different from the conventional compression methods, compressive sampling (CS) provides a new perspective to compress huge data with low energy consumption. Unfortunately, the recordings of underwater acoustic signal (UAS) are nonsparse in time domain. Hence, the current CS methods cannot be used directly for compression and reconstruction of UAS. This study adopts the wavelet-transform-based dictionary matrix to build a framework for sparse representation; then, introduces an approach based on structured approximation $l_0$ (SAL0) norm, which is designed by exploring and exploiting the correlation structure of UAS. The proposed method searches the optimal sparse solution via steepest descent method and then projects the solution to its feasible set. Combing with the compression matrix and dictionary matrix, the estimation of SAL0 method is used for reconstructing nonsparse UAS. Experimental results confirm its recovery quality for practical use, and a clear improvement than the traditional methods including $l_1$ (L1) and orthogonal matching pursuit methods.

Toward a Blind Quality Predictor for Screen Content Images

Blind quality assessment of screen content images (SCIs) is much challenging than traditional natural images. In this paper, we propose a blind quality predictor for SCIs to explore the issue from the perspective of sparse representation. Specifically, we conduct local sparse representation for the textual and pictorial regions, respectively, and conduct global sparse representation for the global SCIs. Subsequently, the underlying relationship between the feature and quality vectors is bridged in a universal sparse representation framework. The quality pooling is comparatively simply that only need to estimate the local and global quality scores and combine them to a total one. Our experimental results show that the proposed predictor can achieve better prediction performance to be in line with subjective assessment.

A Cortical Folding Pattern-Guided Model of Intrinsic Functional Brain Networks in Emotion Processing.

There have been increasing studies demonstrating that emotion processing in humans is realized by the interaction within or among the large-scale intrinsic functional brain networks. Identifying those meaningful intrinsic functional networks based on task-based functional magnetic resonance imaging (task fMRI) with specific emotional stimuli and responses, and exploring the underlying functional working mechanisms of interregional neural communication within the intrinsic functional networks are thus of great importance to understand the neural basis of emotion processing. In this paper, we propose a novel cortical folding pattern-guided model of intrinsic networks in emotion processing: gyri serve as global functional connection centers that perform interregional neural communication among distinct regions via long distance dense axonal fibers, and sulci serve as local functional units that directly communicate with neighboring gyri via short distance fibers and indirectly communicate with other distinct regions via the neighboring gyri. We test the proposed model by adopting a computational framework of dictionary learning and sparse representation of emotion task fMRI data of 68 subjects in the publicly released Human Connectome Project. The proposed model provides novel insights of functional mechanisms in emotion processing.

PolSAR image classification via multimodal sparse representation-based feature fusion

ABSTRACTIt is well known that various features extraction approaches are utilized in polarimetric synthetic aperture (PolSAR) terrain classification for representing the data characteristic. It needs relevant and effective feature fusion algorithms to process complicated features. To address this issue, this article presents a multimodal sparse representation (MSR) framework based algorithm to fuse the different feature vectors from the complicated data space. Polarimetric data features, decomposition features, and the texture features from Pauli colour-coded image are selected to represent multimodal data in different observation modes. The corresponding multimodal manifold regularizations are added to MSR framework to approximate the data structure. Considering the independence and correlation of features, the intrinsic affinity matrices are calculated from this framework. They are processed via local preserve projection algorithm to project the multimodal features into a low dimensionally intrinsic feature space for subsequent classification. Three datasets are utilized in experiments, Western Xi’an, Flevoland, and San Francisco Bay datasets from the Radarsat-2 system in C-band. The effect of regularization parameters and different dimensional fused features are analysed in visualization and quantitation performance. The experiment results demonstrate that the effectiveness and validity of proposed method are superior to other state-of-art methods.

Robust Single-Image Super-Resolution Based on Adaptive Edge-Preserving Smoothing Regularization.

Single-image super-resolution (SR) reconstruction via sparse representation has recently attracted broad interest. It is known that a low-resolution (LR) image is susceptible to noise or blur due to the degradation of the observed image, which would lead to a poor SR performance. In this paper, we propose a novel robust edge-preserving smoothing SR (REPS-SR) method in the framework of sparse representation. An EPS regularization term is designed based on gradient-domain-guided filtering to preserve image edges and reduce noise in the reconstructed image. Furthermore, a smoothing-aware factor adaptively determined by the estimation of the noise level of LR images without manual interference is presented to obtain an optimal balance between the data fidelity term and the proposed EPS regularization term. An iterative shrinkage algorithm is used to obtain the SR image results for LR images. The proposed adaptive smoothing-aware scheme makes our method robust to different levels of noise. Experimental results indicate that the proposed method can preserve image edges and reduce noise and outperforms the current state-of-the-art methods for noisy images.

Multiple 3-D Feature Fusion Framework for Hyperspectral Image Classification

Due to the 3-D nature of hyperspectral images, as well as the spatial properties (such as regularity and continuity) of land covers, many 3-D feature extraction operators have been designed to fully exploit the joint spatial–spectral information. However, the large amount of obtained features can suffer from the “curse of dimensionality” problem, especially for the small training sample set. Moreover, various spatial–spectral features can represent the characteristics of the hyperspectral image from different aspects. In this paper, a multiple 3-D feature fusion framework (M3DF3) has been proposed for hyperspectral image classification. First, we extend the 2-D Gabor surface feature into 3-D (3DSF) domains to comply with the spatial–spectral structure of the hyperspectral image, which is directly applied on the original hyperspectral image instead of the Gabor features. Second, three 3-D feature extraction methods, including the 3-D morphological profile, the 3-D local binary pattern, and the proposed 3DSF, that, respectively, characterize the hyperspectral image from three different angles, i.e., morphology, local dependence, and shape smoothness, are fused under a multitask sparse representation framework to take full advantage of the multiple 3-D features together. The proposed M3DF3 approach was fully tested on three real-world hyperspectral image data, i.e., the widely used Indian Pines, Pavia University, and Houston University. The results show that our method can achieve as high as 68.22%, 79.44%, and 72.84% accuracies, respectively, even when only few samples, i.e., three samples per class, are used for training.

Learning Sparse Representation With Variational Auto-Encoder for Anomaly Detection

Anomaly detection has a wide range of applications in security area such as network monitoring and smart city/campus construction. It has become an active research issue of great concern in recent years. However, most algorithms of the existing studies are powerless for large-scale and high-dimensional data, and the intermediate data extracted by some methods that can handle high-dimensional data will consume lots of storage space. In this paper, we propose a novel sparse representation framework that learns dictionaries based on the latent space of variational auto-encoder. For large-scale data sets, it can play the role of dimensionality reduction to obtain hidden information, and extract more high-level features than hand-crafted features. At the same time, for the storage of normal information, the space cost can be greatly reduced. To verify the versatility and performance of the proposed learning algorithm, we have experimented on different types of anomaly detection tasks, including KDD-CUP data set for network intrusion detection, Mnist data set for image anomaly detection, and UCSD pedestrian’s data set for abnormal event detection in surveillance videos. The experimental results demonstrate that the proposed algorithm outperforms competing algorithms in all kinds of anomaly detection tasks.

Learning Common and Feature-Specific Patterns: A Novel Multiple-Sparse-Representation-Based Tracker.

The use of multiple features has been shown to be an effective strategy for visual tracking because of their complementary contributions to appearance modeling. The key problem is how to learn a fused representation from multiple features for appearance modeling. Different features extracted from the same object should share some commonalities in their representations while each feature should also have some feature-specific representation patterns which reflect its complementarity in appearance modeling. Different from existing multi-feature sparse trackers which only consider the commonalities among the sparsity patterns of multiple features, this paper proposes a novel multiple sparse representation framework for visual tracking which jointly exploits the shared and feature-specific properties of different features by decomposing multiple sparsity patterns. Moreover, we introduce a novel online multiple metric learning to efficiently and adaptively incorporate the appearance proximity constraint, which ensures that the learned commonalities of multiple features are more representative. Experimental results on tracking benchmark videos and other challenging videos demonstrate the effectiveness of the proposed tracker.

DEM generation from contours and a low-resolution DEM

A digital elevation model (DEM) is a virtual representation of topography, where the terrain is established by the three-dimensional co-ordinates. In the framework of sparse representation, this paper investigates DEM generation from contours. Since contours are usually sparsely distributed and closely related in space, sparse spatial regularization (SSR) is enforced on them. In order to make up for the lack of spatial information, another lower spatial resolution DEM from the same geographical area is introduced. In this way, the sparse representation implements the spatial constraints in the contours and extracts the complementary information from the auxiliary DEM. Furthermore, the proposed method integrates the advantage of the unbiased estimation of kriging. For brevity, the proposed method is called the kriging and sparse spatial regularization (KSSR) method. The performance of the proposed KSSR method is demonstrated by experiments in Shuttle Radar Topography Mission (SRTM) 30 m DEM and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) 30 m global digital elevation model (GDEM) generation from the corresponding contours and a 90 m DEM. The experiments confirm that the proposed KSSR method outperforms the traditional kriging and SSR methods, and it can be successfully used for DEM generation from contours.

Nonconvex Weighted $\ell _p$ Minimization Based Group Sparse Representation Framework for Image Denoising

Nonlocal image representation or group sparsity has attracted considerable interest in various low-level vision tasks and has led to several state-of-the-art image denoising techniques, such as BM3D, learned simultaneous sparse coding. In the past, convex optimization with sparsity-promoting convex regularization was usually regarded as a standard scheme for estimating sparse signals in noise. However, using convex regularization cannot still obtain the correct sparsity solution under some practical problems including image inverse problems. In this letter, we propose a nonconvex weighted ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> minimization based group sparse representation framework for image denoising. To make the proposed scheme tractable and robust, the generalized soft-thresholding algorithm is adopted to solve the nonconvex ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> minimization problem. In addition, to improve the accuracy of the nonlocal similar patch selection, an adaptive patch search scheme is proposed. Experimental results demonstrate that the proposed approach not only outperforms many state-of-the-art denoising methods such as BM3D and weighted nuclear norm minimization, but also results in a competitive speed.

An efficient superpixel-based sparse representation framework for hyperspectral image classification

As a powerful classifier, sparse representation-based classification (SRC) has successfully been applied in various visual recognition problems. However, due to the highly correlated bands and insufficient training samples of hyperspectral image (HSI) data, it still remains a challenging problem to effectively apply SRC in HSI. Considering the rich information of spatial structure of materials in HSI, that means the adjacent pixels belong to the same class with a high probability, in this paper, we propose an efficient superpixel-based sparse representation framework for HSI classification. Each superpixel can be regarded as a small region consisting of a number of pixels with similar spectral characteristics. The proposed framework utilizes superpixel to exploit spatial information which can greatly improve classification accuracy. Specifically, SRC is firstly used to classify the HSI data. Meanwhile, an efficient segmentation algorithm is applied to divide the HSI into many disjoint superpixels. Then, each superpixel is used to fuse the SRC classification results in superpixel level. Experimental results on two real-world HSI data sets have shown that the proposed superpixel-based SRC (SP-SRC) framework has a significant improvement over the pixel-based SRC method.

Range profile target recognition using sparse representation based on feature space

A novel method is presented to improve the recognition rate of warhead in this paper. Firstly, a tool for electromagnetic calculation, like CST Microwave Studio, is used to simulate the frequency response of the electromagnetic scattering. Secondly, the echo and further the range profile are acquired from the frequency response by further processing. Thirdly, a set of discriminative features is extracted from the range profiles of the target. Fourthly, these features are used to construct a dictionary for the sparse representation classifier. Finally, the sample of the target can be classified by solving the sparsest coefficients. Since the reconstruction result is determined by a linear combination of the training samples, this method has a good robustness for the variable features. By formulating the problem within a feature-based sparse representation framework, the presented method combines the discriminative features of each sample during the sparse recovery process rather than in a postprocessing manner. Moreover, based on the feature representation space rather than a single feature or image pixel, the constructed dictionary exhibits both strong expressive and discriminative powers that can enhance the classification performance of the test sample. A series of test results based on the simulated data demonstrates the effectiveness of our method.

Local and global sparse representation for no-reference quality assessment of stereoscopic images

No-reference/blind quality assessment of stereoscopic 3D images is much more challenging than 2D images due to the poor understanding of binocular vision. In this paper, we propose a BLind Quality Evaluator for stereoscopic 3D images by learning Local and Global Sparse Representations (BLQELGSR). Specifically, at the training stage, we first construct a large-scale training set by simulating some common distortions that are likely encountered by stereoscopic images, and propose a multi-modal sparse representation framework to characterize the relationship between the feature and quality spaces for all sources of information from left, right and cyclopean views in local and global manners. At the testing stage, based on the derived 3D quality prediction framework, the local and global quality scores from different sources are predicted and combined to drive a final 3D quality score. Experimental results on three 3D image quality databases show that in comparison with the existing methods, the devised BLQELGSR can achieve better prediction performance to be in line with subjective assessment.

Robust adaptive beamforming based on sparse representation technique

The problem of robust adaptive beamforming is addressed within the sparse representation framework. The basic idea of the proposed method is to calculate the adaptive beamformer (BF) with the combination of some easily obtained basic BFs, i.e. the conventional data-independent BFs and the loaded sample matrix inversion BFs. Through using the prior information of the spatial angular sector in which the signal of interest is located, a set of basic BFs pointed at this angular sector are calculated firstly. Then based on the observation that an adaptive BF with favourable performance can be obtained by the combination of only several basic BFs, a new sparse representation-based optimisation model is proposed to search for the adaptive BF. However, the initial optimisation model involves a non-convex constraint which makes the problem intractable. The authors show that the non-convex constraint can be relaxed properly and replaced with a convex one, and the resulting problem can be solved effectively with the interior point method. The obtained BF is robust against model mismatch caused by look direction error, imperfect array calibration etc. The effectiveness and robustness of the proposed method are demonstrated through extensive numerical experiments.

Spatial-Spectral Graph Regularized Kernel Sparse Representation for Hyperspectral Image Classification

This paper presents a spatial-spectral method for hyperspectral image classification in the regularization framework of kernel sparse representation. First, two spatial-spectral constraint terms are appended to the sparse recovery model of kernel sparse representation. The first one is a graph-based spatially-smooth constraint which is utilized to describe the contextual information of hyperspectral images. The second one is a spatial location constraint, which is exploited to incorporate the prior knowledge of the location information of training pixels. Then, an efficient alternating direction method of multipliers is developed to solve the corresponding minimization problem. At last, the recovered sparse coefficient vectors are used to determine the labels of test pixels. Experimental results carried out on three real hyperspectral images point out the effectiveness of the proposed method.

Structured dictionary learning for abnormal event detection in crowded scenes

Abnormal event detection is now a widely concerned research topic, especially for crowded scenes. In recent years, many dictionary learning algorithms have been developed to learn normal event regularities, and have presented promising performance for abnormal event detection. However, they seldom consider the structural information, which plays important roles in many computer vision tasks, such as image denoising and segmentation. In this paper, structural information is explored within a sparse representation framework. On the one hand, we introduce a new concept named reference event, which indicates the potential event patterns in normal video events. Compared with abnormal events, normal ones are more likely to approximate these reference events. On the other hand, a smoothness regularization is constructed to describe the relationships among video events. The relationships consist of both similarities in the feature space and relative positions in the video sequences. In this case, video events related to each other are more likely to possess similar representations. The structured dictionary and sparse representation coefficients are optimized through an iterative updating strategy. In the testing phase, abnormal events are identified as samples which cannot be well represented using the learned dictionary. Extensive experiments and comparisons with state-of-the-art algorithms have been conducted to prove the effectiveness of the proposed algorithm.

Learning Sparse Representation for No-Reference Quality Assessment of Multiply Distorted Stereoscopic Images

Binocular combination under different distortion types poses a great challenge to three-dimensional image quality assessment (3D-IQA). However, the research works on 3D-IQA with multiple distortion types are very limited. In this paper, we first construct a new multiply distorted stereoscopic image database (NBU-MDSID), which is composed of 270 multiply distorted stereoscopic images and 90 singly distorted stereoscopic images that are corrupted simultaneously and independently by blurring, JPEG compression, and noise injection. We then propose a new multimodal blind metric for quality assessment of multiply distorted stereoscopic images. Inspired by multimodal sparse representation framework, modality-specific dictionaries and the corresponding projection matrices are learned from the singly distorted training database at the training stage, and the testing stage only needs to estimate the quality score based on the reconstruction errors. Experimental results demonstrate the effectiveness of our blind metric.

A sparse direction-of-arrival estimation algorithm for MIMO radar in the presence of gain-phase errors

In this paper, the problem of direction-of-arrival (DOA) estimation for monostatic multiple-input multiple-output (MIMO) radar with gain-phase errors is addressed, by using a sparse DOA estimation algorithm with fourth-order cumulants (FOC) based error matrix estimation. Useful cumulants are designed and extracted to estimate the gain and the phase errors in the transmit array and the receive array, thus a reliable error matrix is obtained. Then the proposed algorithm reduces the gain-phase error matrix to a low dimensional one. Finally, with the updated gain-phase error matrix, the FOC-based reweighted sparse representation framework is introduced to achieve accurate DOA estimation. Thanks to the fourth-order cumulants based gain-phase error matrix estimation, and the reweighted sparse representation framework, the proposed algorithm performs well for both white and colored Gaussian noises, and provides higher angular resolution and better angle estimation performance than reduced-dimension MUSIC (RD-MUSIC), adaptive sparse representation (adaptive-SR) and ESPRIT-based algorithms. Simulation results verify the effectiveness and advantages of the proposed method.

Simultaneous denoising and interpolation of 2D seismic data using data-driven non-negative dictionary learning

As a major concern, the existence of unwanted energy and missing traces in seismic data acquisition can degrade interpretation of such data after processing. Instead of analytical dictionaries, data-driven dictionary learning (DDL) methods as a flexible framework for sparse representation, are dedicated to the problem of denoising and interpolation. Due to their meaningful geometric repetitive structures, seismic data are intrinsically low-rank in the time-space domain. On the other hand, noise and missing traces increase the rank of the noisy data. Therefore, the clean data, unlike noise and missing traces, can be modeled as a linear combination of a few elements from a learned dictionary. In this paper, a parts-based 2D DDL scheme is introduced and evaluated for simultaneous denoising and interpolation of seismic data. A special case of versatile non-negative matrix factorization (VNMF) is used to learn a dictionary. In VNMF, smoothness constraint can improve interpolation, and sparse coding helps improving denoising. The proposed method is tested on synthetic and real-field seismic data for simultaneous denoising and interpolation. Through experimental results, the proposed method is determined to be an effective and robust tool that preserves significant components of the signal. Comparison with four state-of-the-art methods further verifies its superior performance.