Search

Due to the complementary properties of different types of sensors, change detection between heterogeneous images receives increasing attention from researchers. However, change detection cannot be handled by directly comparing two heterogeneous images since they demonstrate different image appearances and statistics. In this article, we propose a deep pyramid feature learning network (DPFL-Net) for change detection, especially between heterogeneous images. DPFL-Net can learn a series of hierarchical features in an unsupervised fashion, containing both spatial details and multiscale contextual information. The learned pyramid features from two input images make unchanged pixels matched exactly and changed ones dissimilar and after transformed into the same space for each scale successively. We further propose fusion blocks to aggregate multiscale difference images (DIs), generating an enhanced DI with strong separability. Based on the enhanced DI, unchanged areas are predicted and used to train DPFL-Net in the next iteration. In this article, pyramid features and unchanged areas are updated alternately, leading to an unsupervised change detection method. In the feature transformation process, local consistency is introduced to constrain the learned pyramid features, modeling the correlations between the neighboring pixels and reducing the false alarms. Experimental results demonstrate that the proposed approach achieves superior or at least comparable results to the existing state-of-the-art change detection methods in both homogeneous and heterogeneous cases. [ABSTRACT FROM AUTHOR]

Video object detection, a basic task in the computer vision field, is rapidly evolving and widely used. In recent years, deep learning methods have rapidly become widespread in the field of video object detection, achieving excellent results compared with those of traditional methods. However, the presence of duplicate information and abundant spatiotemporal information in video data poses a serious challenge to video object detection. Therefore, in recent years, many scholars have investigated deep learning detection algorithms in the context of video data and have achieved remarkable results. Considering the wide range of applications, a comprehensive review of the research related to video object detection is both a necessary and challenging task. This survey attempts to link and systematize the latest cutting-edge research on video object detection with the goal of classifying and analyzing video detection algorithms based on specific representative models. The differences and connections between video object detection and similar tasks are systematically demonstrated, and the evaluation metrics and video detection performance of nearly 40 models on two data sets are presented. Finally, the various applications and challenges facing video object detection are discussed. [ABSTRACT FROM AUTHOR]

Global average pooling (GAP) allows convolutional neural networks (CNNs) to localize discriminative information for recognition using only image-level labels. While GAP helps CNNs to attend to the most discriminative features of an object, e.g., head of a bird or one man’s bag, it may suffer if that information is missing due to camera viewpoint changes and intraclass variations in some tasks. To circumvent this issue, we propose one new module to help CNNs to see more, namely, Spatial Rescaling (SpaRs) layer. It introduces spatial relations among the feature map activations back to the model, guiding the model to focus on a broad area in the feature map. With simple implementation, it can be inserted into CNNs of various architectures directly. SpaRs layer consistently improves the performance over the reidentification (re-ID) models. Besides, the new module based on different normalization methods also demonstrates the superiority of fine-grained and general image classification benchmarks. The visualization method shows the changes in activated regions when equipped with the SpaRs layer for better understanding. Our code is publicly available at https://github.com/HRanWang/Spatial-Re-Scaling. [ABSTRACT FROM AUTHOR]

Extracting effective features is always a challenging problem for texture classification because of the uncertainty of scales and the clutter of textural patterns. For texture classification, spectral analysis is traditionally employed in the frequency domain. Recent studies have shown the potential of convolutional neural networks (CNNs) when dealing with the texture classification task in the spatial domain. In this article, we try combining both approaches in different domains for more abundant information and proposed a novel network architecture named contourlet CNN (C-CNN). The network aims to learn sparse and effective feature representations for images. First, the contourlet transform is applied to get the spectral features from an image. Second, the spatial–spectral feature fusion strategy is designed to incorporate the spectral features into CNN architecture. Third, the statistical features are integrated into the network by the statistical feature fusion. Finally, the results are obtained by classifying the fusion features. We also investigated the behavior of the parameters in contourlet decomposition. Experiments on the widely used three texture data sets (kth-tips2-b, DTD, and CUReT) and five remote sensing data sets (UCM, WHU-RS, AID, RSSCN7, and NWPU-RESISC45) demonstrate that the proposed approach outperforms several well-known classification methods in terms of classification accuracy with fewer trainable parameters. [ABSTRACT FROM AUTHOR]

Rapid development of remote sensing (RS) imaging technology makes the acquired images have larger size, higher resolution, and more complex structure, which goes beyond the reach of classical hand-crafted feature-based matching. In this paper, we propose a feature learning approach based on two-branch networks to transform the image matching task into a two-class classification problem. To match two key points, two image patches centered at the key points are entered into the proposed network. The network aims to learn discriminative feature representations for patch matching, so that more matching pairs can be obtained on the premise of maintaining higher subpixel matching accuracy. The proposed network adopts a two-stage training mode to deal with the complex characteristics of RS images. An adaptive sample selection strategy is proposed to determine the size of each patch by the scale of its central key point. Thus, each patch can preserve the texture structure around its key point rather than all patches have a predetermined size. In the matching prediction stage, two strategies, namely, superpixel-based sample graded strategy and superpixel-based ordered spatial matching, are designed to improve the matching efficiency and matching accuracy, respectively. The experimental results and theoretical analysis demonstrate the feasibility, robustness, and effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Based on a generative adversarial network (GAN), a novel version named Task-Oriented GAN is proposed to tackle difficulties in PolSAR image interpretation, including PolSAR data analysis and small sample problem. Besides two typical parts in GAN, i.e., generator (G-Net) and discriminator (D-Net), there is a third part named TaskNet (T-Net) in the Task-Oriented GAN, where T-Net is employed to accomplish a certain task. Two tasks, PolSAR image classification and clustering, are studied in this paper, where T-Net acts as a Classifier and a Clusterer, respectively. The learning procedure of Task-Oriented GAN consists of two main stages. In the first stage, G-Net and D-Net vie with each other like that in a general GAN; in the second stage, G-Net is adjusted and oriented by T-Net so that more samples, which are benefit for the task and called fake data, are generated. As a result, Task-Oriented GAN not only has the advantage of GAN (no-assumption data modeling) but also overcomes the disadvantage of GAN (task-free). After learning, fake data are employed to enrich training set and avoid overfitting; so Task-Oriented GAN performs well even if the manual-labeled data are small. To verify the effectiveness of T-Net, a visualized comparison is provided, where some fake digits generated from Task-Oriented GAN are illustrated along with that from GAN. What is more, considering that there is a great difference between PolSAR data and general data, in our PolSAR image classification and clustering tasks, the specific PolSAR information is inserted into the structure of the Task-Oriented GAN. This enables researchers to mine inherent information in PolSAR data without any data hypothesis and find ways for small sample problem at the same time. Experiment results tested on three PolSAR images show that the proposed method performs well in dealing with PolSAR image classification and clustering. [ABSTRACT FROM AUTHOR]

To detect changed areas in multitemporal polarimetric synthetic aperture radar (SAR) images, this paper presents a novel version of convolutional neural network (CNN), which is named local restricted CNN (LRCNN). CNN with only convolutional layers is employed for change detection first, and then LRCNN is formed by imposing a spatial constraint called local restriction on the output layer of CNN. In the training of CNN/LRCNN, the polarimetric property of SAR image is fully used instead of manual labeled pixels. As a preparation, a similarity measure for polarimetric SAR data is proposed, and several layered difference images (LDIs) of polarimetric SAR images are produced. Next, the LDIs are transformed into discriminative enhanced LDIs (DELDIs). CNN/LRCNN is trained to model these DELDIs by a regression pretraining, and then a classification fine-tuning is conducted with some pseudolabeled pixels obtained from DELDIs. Finally, the change detection result showing changed areas is directly generated from the output of the trained CNN/LRCNN. The relation of LRCNN to the traditional way for change detection is also discussed to illustrate our method from an overall point of view. Tested on one simulated data set and two real data sets, the effectiveness of LRCNN is certified and it outperforms various traditional algorithms. In fact, the experimental results demonstrate that the proposed LRCNN for change detection not only recognizes different types of changed/unchanged data, but also ensures noise insensitivity without losing details in changed areas. [ABSTRACT FROM AUTHOR]

Abstract The increasing number of users participating in location-based social networks has resulted in an information overload problem. Recommendation is a process that can free users from this dilemma. Most algorithms either ignore geographical and social properties, or require a tunable coefficient to determine the effect of each property on the outcome. Simultaneously combining these properties has proven to be a challenge. In this paper, we propose a two-step personalized location recommendation that is based on a multi-objective immune algorithm. It can simultaneously optimize the matching qualities of similarity and geographic properties as two functions, thereby providing location recommendations by improving one desired objective without detracting from the other. In the process, each list provides a different compromise between the similarity of check-in preferences and the geographical influence of the user. The user is offered choices from a set of lists that are compiled from the individual's selection of the various tradeoffs. The advantage of this algorithm is that it can recommend user lists without the need to tune any of the weighting coefficients. Experiments performed using the actual data demonstrated that the proposed algorithm is promising and is an effective means for providing accurate recommendations for a user's desired location. [ABSTRACT FROM AUTHOR]

In this paper, an improved quantum-behaved particle swarm optimization (CL-QPSO), which adopts a new collaborative learning strategy to generate local attractors for particles, is proposed to solve nonlinear numerical problems. Local attractors, which directly determine the convergence behavior of particles, play an important role in quantum-behaved particle swarm optimization (QPSO). In order to get a promising and efficient local attractor for each particle, a collaborative learning strategy is introduced to generate local attractors in the proposed algorithm. Collaborative learning strategy consists of two operators, namely orthogonal operator and comparison operator. For each particle, orthogonal operator is used to discover the useful information that lies in its personal and global best positions, while comparison operator is used to enhance the particle’s ability of jumping out of local optima. By using a probability parameter, the two operators cooperate with each other to generate local attractors for particles. A comprehensive comparison of CL-QPSO with some state-of-the-art evolutionary algorithms on nonlinear numeric optimization functions demonstrates the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

With the vigorous development of big data, the task of large-scale community structure detection is more challenging. In this paper, a large-scale community detection method based on core node and layer-by-layer label propagation is proposed, which is further extended to the detection of overlapping community structures. First, the core nodes whose node degree is greater than the average degree in the graph are found to make effective use of the feature that the core node is the potential community center. This will also avoid the impact of nodes with low node degrees on community structure detection. Then, starting from the core node, label propagation is carried out layer-by-layer according to the node degree and node connection, which effectively improves the accuracy of community detection. The node labels after label propagation are calibrated according to the current attraction of the community to the nodes, which effectively improves the situation of misclassification in the early community detection. Finally, overlapping community detection is carried out based on the non-overlapping community structure. At this time, the overlapping community detection result is more accurate and interpretable. The community detection results on 3 synthetic networks and 12 real datasets show that the proposed algorithm has more advantages than four non-overlapping community detection methods and two overlapping community detection methods. [ABSTRACT FROM AUTHOR]

This paper proposes a dynamic-context cooperative quantum-behaved particle swarm optimization algorithm. The proposed algorithm incorporates a new method for dynamically updating the context vector each time it completes a cooperation operation with other particles. We first explain how this leads to enhanced search ability and improved optimization over previous methods, and demonstrate this empirically with comparative experiments using benchmark test functions. We then demonstrate a practical application of the proposed method, by showing how it can be applied to optimize the parameters for Otsu image segmentation for processing medical images. Comparative experimental results show that the proposed method outperforms other state-of-the-art methods from the literature. [ABSTRACT FROM AUTHOR]

Abstract: The existence of infeasible solutions makes it very difficult to handle constrained optimization problems (COPs) in a way that ensures efficient, optimal and constraint-satisfying convergence. Although further optimization from feasible solutions will typically lead in a direction that generates further feasible solutions, certain infeasible solutions can also provide useful information about the optimal direction of improvement for the objective function. How well an algorithm makes use of these two solutions determines its performance on COPs. This paper proposes a novel selection evolutionary strategy (NSES) for constrained optimization. A self-adaptive selection method is introduced to exploit both informative infeasible and feasible solutions from a perspective of combining feasibility with multi-objective problem (MOP) techniques. Since the global optimal solution of a COP is a feasible non-dominated solution, both non-dominated solutions with low constraint violation and feasible ones with low objective values are beneficial to an evolution process. Thus, the exploration and exploitation of both of these two kinds of solutions are preferred during the selection procedure. Several theorems and properties are given to prove the above assertion. Furthermore, the performance of our method is evaluated using 22 well-known benchmark functions. Experimental results show that the proposed method outperforms state-of-the-art algorithms in terms of the speed of finding feasible solutions and the stability of converging to global optimal solutions. In particular, when dealing with problems that have zero feasibility ratios and more than one active constraint, our method provides feasible solutions within fewer fitness evaluations (FES) and converges to the optimal solutions more reliably than other popular methods from the literature. [Copyright &y& Elsevier]

Abstract: In networked data, linked objects tend to belong to the same class, and densely linked subgraphs are often available. Based on these facts, this paper presents a regularization framework that consists of fitting and regularization terms for transductive learning in networked data. The desirable value of the fitting term is related to the number of labeled data, whereas that of the regularization term is dependent on the structure of the graph. The ratio of these two desirable values is essential for the estimation of the optimal regularization parameters, such as that proposed in our paper. Under the proposed regularization framework, an effective classification algorithm is developed. Two methods are also introduced to incorporate contents of objects into the proposed framework to ultimately improve classification accuracy. Promising experimental results are reported on a toy problem and a paper classification task. [Copyright &y& Elsevier]

Abstract: The generalized fuzzy c-means clustering algorithm with improved fuzzy partition (GFCM) is a novel modified version of the fuzzy c-means clustering algorithm (FCM). GFCM under appropriate parameters can converge more rapidly than FCM. However, it is found that GFCM is sensitive to noise in gray images. In order to overcome GFCMʼs sensitivity to noise in the image, a kernel version of GFCM with spatial information is proposed. In this method, first a term about the spatial constraints derived from the image is introduced into the objective function of GFCM, and then the kernel induced distance is adopted to substitute the Euclidean distance in the new objective function. Experimental results show that the proposed method behaves well in segmentation performance and convergence speed for gray images corrupted by noise. [Copyright &y& Elsevier]

Abstract: Inspired by the principle of gene transposon proposed by Barbara McClintock, a new immune computing algorithm for automatic clustering named as Gene Transposon based Clone Selection Algorithm (GTCSA) is proposed in this paper. The proposed algorithm does not require a prior knowledge of the number of clusters; an improved variant of the clonal selection algorithm is used to determine the satisfied number of clusters and the appropriate partitioning of the data set as well. In addition, a novel operation called antibody gene transposon is introduced to the framework of clonal selection algorithm which can realize to find the satisfied number of cluster automatically. The proposed method has been extensively compared with iterated local search approach (ILS) and three well-known automatic clustering algorithms, including automatic clustering using an improved differential evolution algorithm (ACDE); variable-string-length genetic algorithm based clustering techniques (VGA) and the dynamic clustering approach based on particle swarm optimization (DCPSO). 23 datasets with widely varying characteristics are used to demonstrate the superiority of the GTCSA. In addition, GTCSA is applied to a real world application, namely natural image segmentation, with a good performance obtained. [Copyright &y& Elsevier]

Abstract: Artificial immune systems (AIS) are the computational systems inspired by the principles and processes of the vertebrate immune system. AIS-based algorithms typically mimic the human immune system’s characteristics of learning and adaptability to solve some complicated problems. Here, an artificial immune multi-objective optimization framework is formulated and applied to synthetic aperture radar (SAR) image segmentation. The important innovations of the framework are listed as follows: (1) an efficient and robust immune, multi-objective optimization algorithm is proposed, which has the features of adaptive rank clones and diversity maintenance by K-nearest-neighbor list; (2) besides, two conflicting, fuzzy clustering validity indices are incorporated into this framework and optimized simultaneously and (3) moreover, an effective, fused feature set for texture representation and discrimination is constructed and researched, which utilizes both the Gabor filter’s ability to precisely extract texture features in low- and mid-frequency components and the gray level co-occurrence probability’s (GLCP) ability to measure information in high-frequency. Two experiments with synthetic texture images and SAR images are implemented to evaluate the performance of the proposed framework in comparison with other five clustering algorithms: fuzzy C-means (FCM), single-objective genetic algorithm (SOGA), self-organizing map (SOM), wavelet-domain hidden Markov models (HMTseg), and spectral clustering ensemble (SCE). Experimental results show the proposed framework has obtained the better performance in segmenting SAR images than other five algorithms and behaves insensitive to the speckle noise. [Copyright &y& Elsevier]

Abstract: Artificial immune systems are a kind of new computational intelligence methods which draw inspiration from the human immune system. Most immune system inspired optimization algorithms are based on the applications of clonal selection and hypermutation, and known as clonal selection algorithms. These clonal selection algorithms simulate the immune response process based on principles of Darwinian evolution by using various forms of hypermutation as variation operators. The generation of new individuals is a form of the trial and error process. It seems very wasteful not to make use of the Baldwin effect in immune system to direct the genotypic changes. In this paper, based on the Baldwin effect, an improved clonal selection algorithm, Baldwinian Clonal Selection Algorithm, termed as BCSA, is proposed to deal with optimization problems. BCSA evolves and improves antibody population by four operators, clonal proliferation, Baldwinian learning, hypermutation, and clonal selection. It is the first time to introduce the Baldwinian learning into artificial immune systems. The Baldwinian learning operator simulates the learning mechanism in immune system by employing information from within the antibody population to alter the search space. It makes use of the exploration performed by the phenotype to facilitate the evolutionary search for good genotypes. In order to validate the effectiveness of BCSA, eight benchmark functions, six rotated functions, six composition functions and a real-world problem, optimal approximation of linear systems are solved by BCSA, successively. Experimental results indicate that BCSA performs very well in solving most of the test problems and is an effective and robust algorithm for optimization. [Copyright &y& Elsevier]

The receiver operating characteristic (ROC) and detection error tradeoff (DET) curves are frequently used in the machine learning community to analyze the performance of binary classifiers. Recently, the convex-hull-based multiobjective genetic programming algorithm was proposed and successfully applied to maximize the convex hull area for binary classification problems by minimizing false positive rate and maximizing true positive rate at the same time using indicator-based evolutionary algorithms. The area under the ROC curve was used for the performance assessment and to guide the search. Here we extend this research and propose two major advancements: Firstly we formulate the algorithm in detection error tradeoff space, minimizing false positives and false negatives, with the advantage that misclassification cost tradeoff can be assessed directly. Secondly, we add complexity as an objective function, which gives rise to a 3D objective space (as opposed to a 2D previous ROC space). A domain specific performance indicator for 3D Pareto front approximations, the volume above DET surface, is introduced, and used to guide the indicator-based evolutionary algorithm to find optimal approximation sets. We assess the performance of the new algorithm on designed theoretical problems with different geometries of Pareto fronts and DET surfaces, and two application-oriented benchmarks: (1) Designing spam filters with low numbers of false rejects, false accepts, and low computational cost using rule ensembles, and (2) finding sparse neural networks for binary classification of test data from the UCI machine learning benchmark. The results show a high performance of the new algorithm as compared to conventional methods for multicriteria optimization. [ABSTRACT FROM AUTHOR]

Recently, the majority of successful matching approaches are based on convolutional neural networks, which focus on learning the invariant and discriminative features for individual image patches based on image content. However, the image patch matching task is essentially to predict the matching relationship of patch pairs, that is, matching (similar) or non-matching (dissimilar). Therefore, we consider that the feature relation (FR) learning is more important than individual feature learning for image patch matching problem. Motivated by this, we propose an element-wise FR learning network for image patch matching, which transforms the image patch matching task into an image relationship-based pattern classification problem and dramatically improves generalization performances on image matching. Meanwhile, the proposed element-wise learning methods encourage full interaction between feature information and can naturally learn FR. Moreover, we propose to aggregate FR from multilevels, which integrates the multiscale FR for more precise matching. Experimental results demonstrate that our proposal achieves superior performances on cross-spectral image patch matching and single spectral image patch matching, and good generalization on image patch retrieval. [ABSTRACT FROM AUTHOR]

• A spatial kernel generation module is proposed. • A spectral kernel generation module is proposed. • A multi-scale spectral channel attention mechanism is proposed. • A strategy for feature fusion through convolution is proposed. • Our network structure has better classification results than the state-of-the-art. With the widespread use of deep learning methods, more and more classification models based on hyperspectral images (HSI) have been continuously proposed. However, due to the characteristics of high dimensionality and low resolution of HSI, the traditional convolutional neural network (CNN) model cannot effectively process it. In this paper, we propose a classification network, called spatial and spectral kernels generation network (SSKNet), to generate spatial and spectral convolution kernels based on image characteristics, and use them to replace the traditional initialization convolution kernels. We can obtain representative kernels with stronger spatial correlation with the region segmentation, clustering, and mapping operations of the spatial kernels generation module (SaKG Module). Simultaneously, we also propose a spectral kernels generation module (SeKG module), which integrates the multi-scale correlation characteristics of different bands into the spectral attention mechanism, making the generated spectral kernels more accurate. Combining the spatial and spectral kernels allows the network to extract salient features and achieve feature fusion efficiently. Experimental results based on multiple HSI data sets show that the proposed method has better classification accuracy and generalization performance than existing methods. [ABSTRACT FROM AUTHOR]

Abstract: Spectral clustering algorithms have been successfully used in the field of pattern recognition and computer vision. The widely used similarity measure for spectral clustering is Gaussian kernel function which measures the similarity between data points. However, it is difficult for spectral clustering to choose the suitable scaling parameter in Gaussian kernel similarity measure. In this paper, utilizing the prototypes and partition matrix obtained by fuzzy c-means clustering algorithm, we develop a fuzzy similarity measure for spectral clustering (FSSC). Furthermore, we introduce the K-nearest neighbor sparse strategy into FSSC and apply the sparse FSSC to texture image segmentation. In our experiments, we firstly perform some experiments on artificial data to verify the efficiency of the proposed fuzzy similarity measure. Then we analyze the parameters sensitivity of our method. Finally, we take self-tuning spectral clustering and Nyström methods for baseline comparisons, and apply these three methods to the synthetic texture and remote sensing image segmentation. The experimental results show that the proposed method is significantly effective and stable. [Copyright &y& Elsevier]

Abstract: In this paper, we present a family of complex-valued support vector classifiers (CSVCs) based on the definition of the complex sign inspired by the modulation in digital communications and the complex-valued kernel functions. We also propose a theorem to construct the complex-valued Mercer kernels and the corresponding kernel function groups. CSVC algorithms include binary (2-state) CSVC (BCSVC), quadrature (4-state) CSVC (QCSVC) and some multi-state CSVCs. In this paper, we focus on QCSVC. For a quadrature complex-valued classification problem, QCSVC is identical to the 4-quadrature amplitude modulation demodulation methods in digital communications. Finally, the simulated experiments confirm the validity and the efficiency of CSVCs. [Copyright &y& Elsevier]

Deep belief network (DBN) is an import deep learning model and restricted Boltzmann machine (RBM) is one of its basic models. The traditional DBN and RBM have numerous redundant features. Hence an improved strategy is required to perform sparse operations on them. Previously, we have proposed our own sparse DBN (SDBN): using a multi-objective optimization (MOP) algorithm to learn sparse features, which solves the contradiction between the reconstruction error and network sparsity of RBM. Due to the optimization algorithm and millions of parameters of the network itself, the training process is difficult. Therefore, in this paper, we propose an efficient parallel strategy to speed up the training of SDBN networks. Self-adaptive Quantum Multi-objectives Evolutionary algorithm based on Decomposition (SA-QMOEA/D) that we have proposed as the multi-objective optimization algorithm has the hidden parallelism of populations. Based on this, we not only parallelize the DBN network but also realize the parallelism of the multi-objective optimization algorithm. In order to further verify the advantages of our approach, we apply it to the problem of facial expression recognition (FER). The obtained experimental results demonstrate that our parallel algorithm achieves a significant speedup performance and a higher accuracy rate over previous CPU implementations and other conventional methods. [ABSTRACT FROM AUTHOR]

Semi-supervised non-negative matrix factorization (NMF) exploits the strengths of NMF in effectively learning local information contained in data and is also able to achieve effective learning when only a small fraction of data is labeled. NMF is particularly useful for dimensionality reduction of high-dimensional data. However, the mapping between the low-dimensional representation, learned by semi-supervised NMF, and the original high-dimensional data contains complex hierarchical and structural information, which is hard to extract by using only single-layer clustering methods. Therefore, in this article, we propose a new deep learning method, called semi-supervised graph regularized deep NMF with bi-orthogonal constraints (SGDNMF). SGDNMF learns a representation from the hidden layers of a deep network for clustering, which contains varied and unknown attributes. Bi-orthogonal constraints on two factor matrices are introduced into our SGDNMF model, which can make the solution unique and improve clustering performance. This improves the effect of dimensionality reduction because it only requires a small fraction of data to be labeled. In addition, SGDNMF incorporates dual-hypergraph Laplacian regularization, which can reinforce high-order relationships in both data and feature spaces and fully retain the intrinsic geometric structure of the original data. This article presents the details of the SGDNMF algorithm, including the objective function and the iterative updating rules. Empirical experiments on four different data sets demonstrate state-of-the-art performance of SGDNMF in comparison with six other prominent algorithms. [ABSTRACT FROM AUTHOR]

In this article, a new deep neural network based on sparse filtering and manifold regularization (DSMR) is proposed for feature extraction and classification of polarimetric synthetic aperture radar (PolSAR) data. DSMR uses a novel deep neural network (DNN) to automatically learn features from raw SAR data. During preprocessing, the spatial information between pixels on PolSAR images is exploited to weight each data sample. Then, in the pretraining and fine-tuning, DSMR uses the population sparsity and the lifetime sparsity (dual sparsity) to learn the global features and preserves the local structure of data by neighborhood-based manifold regularization. The dual sparsity only needs to tune a few parameters, and the manifold regularization cuts down the number of training samples. Experimental results on synthesized and real PolSAR data sets from different SAR systems show that DSMR can improve classification accuracy compared with conventional DNNs, even for data sets with a large angle of incidence. [ABSTRACT FROM AUTHOR]

• A new multi-layer interaction preference based multi-objective optimization algorithm through decomposition (MLIP-MOEA/D) is proposed. • A fast way to get preference region is established, during the selection process, the preference weight vector is redefined using the angle-based method. • A novel multi-layer interaction strategy for searching the accurate solutions is proposed. During evolution, the preferred region is reduced gradually, making it easier for user to get the final preferred solution(s). • Besides ZDT problems with two objectives, the proposed algorithm shows its potentiality for solving many-objective problems, i.e., DTLZ problems with 3, 5, 8, 10 objectives. Many problems in real world have not only one objective to be met. In the majority of cases, a set of trade-off solutions which spread evenly along the entire Pareto optimal front are generated by multi-objective evolutionary algorithms (MOEAs). Taking the preference of decision maker (DM) into consideration, some specified solutions can be obtained, which is of great interest in practical applications. In this paper, a novel multi-layer interaction preference based multi-objective evolutionary algorithm through decomposition (denoted as MLIP-MOEA/D) is proposed. In MLIP-MOEA/D, a multi-layer interactive strategy is developed during evolution, in the first-layer interaction, the DM will provide a reference vector and an initial radius to determine a preference range, then all solutions in this range will be updated. The algorithm will stop if the DM is satisfied with the first output result, otherwise it will go on to the second-layer interaction. In this step, the most preferred solution generated from the first-layer interaction will be chosen as the new preference direction, and the weight vector is redefined by the angle-based method, and the range of preferred region is reduced gradually, until the closest solution that meet the DM's need is found. The algorithm is tested on a set of benchmark problems including DTLZ problems with more than three objectives, the experimental studies show that the proposed algorithm can effectively search the preferred solutions with the preference information and successfully deal with many-objective optimization problems. [ABSTRACT FROM AUTHOR]

Single image super-resolution (SR) reconstruction is an ill-posed inverse problem because the high-resolution (HR) image, obtained from the low-resolution (LR) image, is non-unique or unstable. In this paper, single image SR reconstruction is treated as an optimization problem, and a new single image SR method, based on a genetic algorithm and regularization prior model, is proposed. In the proposed method, the optimization problem is constructed with a regularization prior model which consists of the non-local means (NLMs) filter, total variation (TV) and adaptive sparse domain selection (ASDS) scheme for sparse representation. In order to avoid local optimization, we combine the genetic algorithm and the iterative shrinkage algorithm to deal with the regularization prior model. Compared with several other state-of-the-art algorithms, the proposed method demonstrates better performances in terms of both numerical analysis and visual effect. [ABSTRACT FROM AUTHOR]

High-order graph matching aims at establishing correspondences between two sets of feature points using high-order constraints. It is usually formulated as an NP-hard problem of maximizing an objective function. This paper introduces a discrete particle swarm optimization algorithm for resolving high-order graph matching problems, which incorporates several re-defined operations, a problem-specific initialization method based on heuristic information, and a problem-specific local search procedure. The proposed algorithm is evaluated on both synthetic and real-world datasets. Its outstanding performance is validated in comparison with three state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

This paper presents a novel change detection approach for synthetic aperture radar images based on deep learning. The approach accomplishes the detection of the changed and unchanged areas by designing a deep neural network. The main guideline is to produce a change detection map directly from two images with the trained deep neural network. The method can omit the process of generating a difference image (DI) that shows difference degrees between multitemporal synthetic aperture radar images. Thus, it can avoid the effect of the DI on the change detection results. The learning algorithm for deep architectures includes unsupervised feature learning and supervised fine-tuning to complete classification. The unsupervised feature learning aims at learning the representation of the relationships between the two images. In addition, the supervised fine-tuning aims at learning the concepts of the changed and unchanged pixels. Experiments on real data sets and theoretical analysis indicate the advantages, feasibility, and potential of the proposed method. Moreover, based on the results achieved by various traditional algorithms, respectively, deep learning can further improve the detection performance. [ABSTRACT FROM PUBLISHER]

Social computing is a new paradigm for information and communication technology. Social network analysis is one of the theoretical underpinnings of social computing. Community structure detection is believed to be an effective tool for social network analysis. Uncovering community structures in social networks can be regarded as clustering optimization problems. Because social networks are characterized by dynamics and huge volumes of data, conventional nature-inspired optimization algorithms encounter serious challenges when applied to solving large-scale social network clustering optimization problems. In this study, we put forward a novel particle swarm optimization algorithm to reveal community structures in social networks. The particle statuses are redefined under a discrete scenario. The status updating rules are reconsidered based on the network topology. A greedy strategy is designed to drive particles to a promising region. To this end, a greedy discrete particle swarm optimization framework for large-scale social network clustering is suggested. To determine the performance of the algorithm, extensive experiments on both synthetic and real-world social networks are carried out. We also compare the proposed algorithm with several state-of-the-art network community clustering methods. The experimental results demonstrate that the proposed method is effective and promising for social network clustering. [ABSTRACT FROM AUTHOR]

Image segmentation is an important and well-known ill-posed inverse problem in computer vision. It is a process of assigning a label to each pixel in a digital image so that pixels with the same label have similar characteristics. Chan–Vese model which belongs to partial differential equation approaches has been widely used in image segmentation tasks. Chan–Vese model has to optimize a non-convex problem. It usually converges to local minima. Furthermore, the length penalty item which is critical to the final results of Chan–Vese model makes the model be sensitive to parameter settings and costly in computation. In order to overcome these drawbacks, a novel bi-convex fuzzy variational image segmentation method is proposed. It is unique in two aspects: (1) introducing fuzzy logic to construct a bi-convex object function in order to simplify the procedure of finding global optima and (2) efficiently combining the length penalty item and the numerical remedy method to get better results and to bring robustness to parameter settings and greatly reduce computation costs. Experiments on synthetic, natural, medical and radar images have visually or quantitatively validated the superiorities of the proposed method compared with five state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

Abstract: Capacitated Arc Routing Problem (CARP) has drawn much attention during the last few years. In addition to the goal of minimizing the total cost of all the routes, many real-world applications of CARP also need to balance these routes. The Multi-objective CARP (MO-CARP) commonly exists in practical applications. In order to solve MO-CARP efficiently and accurately, this paper presents a Multi-population Cooperative Coevolutionary Algorithm (MPCCA) for MO-CARP. Firstly, MPCCA applies the divide-and-conquer method to decompose the whole population into multiple subpopulations according to their different direction vectors. These subpopulations evolve separately in each generation and the adjacent subpopulations can share their individuals in the form of cooperative subpopulations. Secondly, multiple subpopulations are used to search different objective subregions simultaneously, so individuals in each subpopulation have a different fitness function, which can be modeled as a Single Objective CARP (SO-CARP). The advanced MAENS approach for single-objective CARP can be used to search each objective subregion. Thirdly, the internal elitism archive is used to construct evolutionary pool for each subregion, which greatly speeds up the convergence. Lastly, the fast nondominated ranking and crowding distance of NSGA-II are used for selecting the offspring and keeping the diversity. MPCCA is tested on 91 CARP benchmarks. The experimental results show that MPCCA obtains better generalization performance over the compared algorithms. [Copyright &y& Elsevier]

Abstract: In this paper we propose a novel Dictionary Learning and Sparse Representation-based Classifier (DLSRC) for image segmentation. In DLSRC, instances-based learning is adopted to find representative dictionaries that can sparsely code various classes of prototype samples in images. Then an incremental version of DLSRC, IDLSRC, is advanced for incremental learning of accumulating knowledge obtained from labeled data. The unsupervised clustering algorithm provides initial labeled samples, and then the labels of candidate samples are incrementally predicted by defining a consistency-enhanced evaluation function. Some experiments are taken on both the artificial texture images and real Synthetic Aperture Radar (SAR) images, to investigate the performance of DLSRC and IDLSRC. Some aspects including (1) the comparison of DLSRC with the Sparse Representation based Classifier (SRC) and some unsupervised clustering approaches, (2) the comparison of IDLSRC with DLSRC, are tested, and the results prove the superiority of our proposed method to its counterparts. [Copyright &y& Elsevier]

Abstract: In this paper, we propose automatic image segmentation using constraint learning and propagation. Recently, kernel learning is receiving much attention because a learned kernel can fit the given data better than a predefined kernel. To effectively learn the constraints generated by initial seeds for image segmentation, we employ kernel propagation (KP) based on kernel learning. The key idea of KP is first to learn a small-sized seed-kernel matrix and then propagate it into a large-sized full-kernel matrix. By applying KP to automatic image segmentation, we design a novel segmentation method to achieve high performance. First, we generate pairwise constraints, i.e., must-link and cannot-link, from initially selected seeds to make the seed-kernel matrix. To select the optimal initial seeds, we utilize global k-means clustering (GKM) and self-tuning spectral clustering (SSC). Next, we propagate the seed-kernel matrix into the full-kernel matrix of the entire image, and thus image segmentation results are obtained. We test our method on the Berkeley segmentation database, and the experimental results demonstrate that the proposed method is very effective in automatic image segmentation. [Copyright &y& Elsevier]

Abstract: Recent results of Compressive Sampling (CS) have demonstrated its feasibility in high-resolution radar targets estimation and imaging [2,10,14,15,17,19,23,29,30,32–34]. However, the signal recovery is reduced to seeking a sparse solution to an underdetermined linear system of equations. It is potentially very difficult because even finding a solution that approximates the true minimum is NP-hard. In this paper, we introduce Memetic Algorithm (MA) to solve this non-convex l 0-norm minimization problem, and design a compressive receiver for high-resolution range-reflectivity estimation of multiple radar targets. A double-population MA is proposed, where the position population is used to evaluate the ranges, and the coefficient population is used to realize a local search of target reflectivities. By combining the global search with a local searching operation to exploit the available knowledge in the recovery, the proposed MA outperforms the general purpose optimization algorithms in terms of the quality of solution. Some experiments are taken to investigate the performance of this compressive receiver at different sampling rates, and the results show the superiority to its counterparts in both noiseless environment and noisy, cluttered environment. [Copyright &y& Elsevier]