The rapid development of the digital age has led to a qualitative leap in social media. To meet the cognitive needs of users, social media platforms have been mining users’ private information and disseminating information through various means. However, these platforms lack effective management of information release and various forms of emotional expressions make public propaganda increasingly diverse and complex. Therefore, accurately identifying the relationships between multimodal data poses a challenge. An effective modal representation must consider both the consistency of multimodal data and the complementarity of single-modal data. However, existing methods focus on fusing different modal features into a unified feature representation, while neglecting to evaluate the reliability of prediction results. In this article, we disentangle the consistency and complementarity in the fused representation problem of multimodal data. We construct the modal private task (unique) by using the Dirichlet distribution and evidence theory to solve the uncertainty of each modal prediction. The model can output the uncertainty of prediction and learn complementary information through the fusion of decision layers. At the same time, we construct the modal common task using a low-rank tensor fusion model to learn consistent features. Finally, we compare the model with the current mainstream methods on three public datasets, and the experimental results show that the performance of our method reaches the level of current advanced algorithms.

Hyperspectral image (HSI) classification is a hot topic in the remote sensing community. However, it is challenging to fully use spatial–spectral information for HSI classification due to the high dimensionality of the data, high intraclass variability, and the limited availability of training samples. To deal with these issues, we propose a novel feature extraction method called nonlocal correntropy matrix (NLCM) representation in this letter. NLCM can characterize the spectral correlation and effectively extract discriminative features for HSI classification. We verify the effectiveness of the proposed method on two widely used datasets. The results show that NLCM performs better than the state-of-the-art methods, especially when the training set size is small. Furthermore, the experimental results also demonstrate that the proposed method outperforms compared methods significantly when the land covers are complex and with irregular distributions.

Recently, Hyperspectral Image Classification (HIC) with noisy labels is attracting increasing interest. However, existing methods usually neglect to explore feature-dependent knowledge to reduce label noise, and thus perform poorly when the noise ratio is high or the clean samples are limited. In this paper, a novel Triple Contrastive Representation Learning (TCRL) framework is proposed from a deep clustering perspective for robust HIC with noisy labels. The TCRL explores the cluster-level, instance-level, and structure-level representation of HIC by defining triple learning loss. First, the strong and weak transformation are defined for hyperspectral data augmentation. Then, a simple yet effective lightweight Spectral Prior Attention-based Network (SPAN) is presented for spatial-spectral feature extraction of all augmented samples. Additionally, cluster-level and instance-level contrastive learning are performed on two projection subspaces for clustering and distinguishing samples respectively. Meanwhile, structure-level representation learning is employed to maximize the consistency of data after different projections. Taking the feature-dependent information learned by triple representation learning, our proposed end-to-end TCRL can effectively alleviate the overfitting of classifier to noisy labels. Extensive experiments have been taken on three public datasets with various noise ratios and two types of noise. The results show that the proposed TCRL could provide more robust classification results when training on noisy datasets compared with state-of-the-art methods, especially when clean samples are limited. The code will be available at https://github.com/Zhangxy1999.

Abstract A new spectral‐spatial hyperspectral image (HSI) classification method called hierarchical broad learning system (HBLS) has been proposed in this paper. Specifically, it combines wavelet, broad learning system (BLS) and Gabor filters into a hierarchical structure. First of all, wavelet is used to reduce the observation noise of HSIs. Then BLS is adopted to acquire a set of pixelwise probability maps from the input data, and Gabor filters are used to explore spatial information by refining these probability maps. These two operations (BLS and Gabor filtering) are alternated to form a hierarchical architecture. And the discriminative spectral‐spatial features can be extracted at each layer of the hierarchical architecture. Finally, the spectral‐spatial features are fed into the standard BLS for classification. Experimental results on three widely used HSIs reveal that HBLS outperforms some state‐of‐the‐art methods in terms of classification accuracy and sample complexity.

The functional principal component analysis (FPCA) method can effectively solve the problems of the high dimensionality of data, large information redundancy, and noise interference in hyperspectral image (HSI) classification. However, this unsupervised FPCA cannot make full use of the label information of training samples or spatial information, so that it is impossible to obtain satisfactory classification results. In this letter, a set of improved FPCA methods for HSI classification are proposed. First, the B-spline basis system is used to establish the functional data fitting model, which can convert discrete spectral information into continuous spectral curves and lay the foundation for functional feature extraction. Second, a supervised FPCA (SFPCA) method is built for extracting more effective functional features by making full use of the label information of training samples. Furthermore, to overcome the lack of training samples, two semisupervised FPCA (SSFPCA) methods are proposed for extracting more discriminative functional features and improving the classification accuracies. Finally, we perform the local mean filtering method on the HSI in order to extract the spatial information for each pixel, and then design spectral-spatial classification frameworks based on improved FPCA. Experiments on the commonly used HSI dataset show that improved FPCA can achieve higher classification accuracies than FPCA, and the proposed functional spectral-spatial classification frameworks can greatly improve classification accuracies.

The hyperspectral images (HSIs) classification technique has received widespread attention in the field of remote sensing. However, how to achieve satisfactory classification performance in the presence of a large amount of noise is still a problem worthy of consideration. In this article, a local correntropy matrix (LCEM)-based spatial–spectral feature representation method is proposed for HSI classification. Motivated by the successful application of information-theoretic learning (ITL), we propose to adopt correntropy matrix to represent the spatial–spectral features of HSI. Specifically, the dimension reduction is first performed on the original hyperspectral data. Then, for each pixel, we select its local neighbors within a sliding window using cosine distance for the construction of the LCEM. In this way, each pixel can be characterized as an LCEM. Finally, all the correntropy matrices are fed into a support vector machine (SVM) for final classification. In addition, we also propose a novel way to determine the size of the local window based on standard deviation. Because the LCEM as the feature descriptor can characterize discriminative spatial–spectral features, the proposed method has shown great interclass separability and intraclass compactness. Compared with other advanced approaches, the proposed LCEM method has achieved competitive performance in both evaluation indexes and visual effects, especially when the training size is very small.

Considering that multiple features can improve the classification performance as they contain diversity information of images, a multiple-feature latent space learning-based method is proposed for hyperspectral image (HSI) classification in this letter. In the proposed method, a latent space that contains diversity information of multiple features and transformation matrices between the latent space and features are both learned. Moreover, spatial information is used for labeling unlabeled samples in the classification. Experimental results on the Indian Pines and University of Pavia data sets demonstrate the effectiveness of the proposed method.

This paper proposes a novel spectral-spatial feature representation method for hyperspectral image (HSI) classification. It combines the advantages of adaptive weighted filtering (AWF) and local covariance matrix representation (L-CMR) to make full use of the spatial similarity and correlation among different spectral bands. Specifically, the proposed method first uses the maximum noise fraction (MNF) to reduce the dimensionality of HSI. Then, multiscale AWF (MAWF) is applied to make use of spatial information. N ext' the spectral-spatial features are obtained by calculating the local covariance matrix of the given pixel and its neighbors. Finally, the learned spectral-spatial features of each pixels are fed into support vector machine (SVM) for classification. Experimental results on two publicly available HSI datasets show that the proposed method is superior to several existing methods in terms of both classification accuracy and classification visual effect, especially when the number of training samples is small.

This article proposes a functional data discriminant analysis (FDDA) method for hyperspectral image (HSI) classification. This method analyzes and processes the HSI data from a functional point of view, which is a novel perspective in HSI processing. The classical methods achieve dimensionality reduction by directly eliminating the redundancy of the HSI data. However, the proposed method extracts the functional features by utilizing the redundancy of the HSI data. Functional features can effectively reveal inherent characteristics of the HSI data with the change in the wavelengths. Based on this, a regularized weighted fitting model is first built for converting a spectral vector into a spectral curve. Second, an FDDA method defined in the function field is presented for extracting the functional features of the spectral curves. Finally, a novel spectral-spatial framework is designed for classification tasks of HSI data sets. Experimental results in three commonly used HSI data sets indicate that the proposed method is effective and leads to promising classification results compared with some benchmarking methods. More importantly, the work tries to diversify and develop the existing theory and methods of HSI classification from discrete (vector) data learning methods to continuous (functional) data learning methods.

Robust and efficient infrared (IR) small target detection plays an important role in image processing for IR remote sensing. In order to detect the IR small target with high detection rate, low false alarm rate (FAR), and high detection speed, a novel method called high-boost-based multiscale local contrast measure (HB-MLCM) is proposed in this letter. First, improved high boost filter is proposed to enhance the high frequency signal where the target may appear and suppress the low frequency signal. Then, a simple MLCM is proposed for further enhancing the target and suppressing the background. Finally, a simple and adaptive thresholding method is used to segment targets from the contrast map. Experimental results on three real image sequences with various typical complex backgrounds demonstrate that the proposed method can effectively detect the target with faster speed, higher detection rate, and lower FAR compared with the state-of-the-art methods.

In this paper, we present a hierarchical feature learning method called Stacked Tensor Subspace Learning (STSL). It can jointly learn spectral and spatial features of hyperspectral images (HSIs) by iteratively abstracting neighboring regions. STSL is able to learn discriminative spectral-spatial features of the input HSI at different scales. In STSL, the joint spectral and spatial features are extracted using Marginal Fisher Analysis (MFA) and Tensor Principal Component Analysis (TPCA). Then Kernel-based Extreme Learning Machine (KELM), a shallow neural network, is embedded in the proposed method to classify image pixels. The important contributions to the success of STSL are exploiting local spatial structure of HSI by using tensor method and designing hierarchical architecture. Extensive experimental results on two challenging HSI data sets taken from the Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) and Reflective Optics System Imaging Spectrometer (ROSIS) airborne sensors show that the proposed method can produce good classification accuracy with smaller training sets.

In the hyperspectral image classification area, a few number of labeled samples is a bottleneck for the improvement of classification accuracy. In order to tackle this problem, multiple one-dimensional embedding interpolation (M1DEI) has been used for hyperspectral image classification and achieved promising results. Despite the success, the complexity of M1DEI prevents its practical application. On the other hand, the percentage of newly added samples is set by experience when enlarging the labeled set. In this paper we develop a method by extending the M1DEI method with local strategy, called multiple local one-dimensional embedding interpolation (ML1DEI). We only map the labeled samples and their local spatial neighbors into the one-dimensional (1D) space. The local strategy can reduce the complexity of M1DEI, since only labeled samples and their neighbors need to be mapped. In addition, the local strategy ensures all these newly labeled samples come from the spatial neighborhood of labeled samples. Then, during the merging stage, we can incorporate all of them with the labeled samples. Moreover, the proposed ML1DEI can incorporate the spatial information and make full use of the unlabeled samples. Compared with other spatial-spectral classification methods, the proposed ML1DEI method obtains promising results. Experimental results on the commonly used hyperspectral data set validate the effectiveness of the proposed method.

This paper proposes a spectral-spatial feature learning (SSFL) method to obtain robust features of hyperspectral images (HSIs). It combines the spectral feature learning and spatial feature learning in a hierarchical fashion. Stacking a set of SSFL units, a deep hierarchical model called the spectral-spatial networks (SSN) is further proposed for HSI classification. SSN can exploit both discriminative spectral and spatial information simultaneously. Specifically, SSN learns useful high-level features by alternating between spectral and spatial feature learning operations. Then, kernel-based extreme learning machine (KELM), a shallow neural network, is embedded in SSN to classify image pixels. Extensive experiments are performed on two benchmark HSI datasets to verify the effectiveness of SSN. Compared with state-of-the-art methods, SSN with a deep hierarchical architecture obtains higher classification accuracy in terms of the overall accuracy, average accuracy, and kappa ( κ ) coefficient of agreement, especially when the number of the training samples is small.

Robust small target detection of low signal-to-noise ratio (SNR) is very important in infrared search and track applications for self-defense or attacks. Consequently, an effective small target detection algorithm inspired by the contrast mechanism of human vision system and derived kernel model is presented in this paper. At the first stage, the local contrast map of the input image is obtained using the proposed local contrast measure which measures the dissimilarity between the current location and its neighborhoods. In this way, target signal enhancement and background clutter suppression are achieved simultaneously. At the second stage, an adaptive threshold is adopted to segment the target. The experiments on two sequences have validated the detection capability of the proposed target detection method. Experimental evaluation results show that our method is simple and effective with respect to detection accuracy. In particular, the proposed method can improve the SNR of the image significantly.

Background suppression is vitally important for the small target detection, which aims to enhance targets and improve the signal‐to‐noise ratio of small target images. Consequently, the study proposes a background suppression approach based on the fast local reverse entropy operator, which is designed according to the fact that the appearance of a small target could result in the great change of the value of local reverse entropy in the local region. The operator is adopted to suppress complex backgrounds of small target images in order to enhance small targets, and then bring about high probabilities of detection and low probabilities of false alarm in the small target detection. Both quantitative and qualitative analyses contribute to confirm the validity and efficiency of the proposed approach.

In this paper, a hierarchical feature extraction method is proposed for image recognition. The key idea of the proposed method is to extract an effective feature, called local neural response (LNR), of the input image with nontrivial discrimination and invariance properties by alternating between local coding and maximum pooling operation. The local coding, which is carried out on the locally linear manifold, can extract the salient feature of image patches and leads to a sparse measure matrix on which maximum pooling is carried out. The maximum pooling operation builds the translation invariance into the model. We also show that other invariant properties, such as rotation and scaling, can be induced by the proposed model. In addition, a template selection algorithm is presented to reduce computational complexity and to improve the discrimination ability of the LNR. Experimental results show that our method is robust to local distortion and clutter compared with state-of-the-art algorithms.

The paper introduces two different types of frequencies of which one is the Arccosine Instantaneous Frequency (ArccosineIF) for the so called axial simple waves (ASWs); and the other is the α-Counting Instantaneous Frequency (α-CIF) for a more general class of signals called simple waves (SWs). The classes ASW and SW contain a wide range of signals for which the concept instantaneous frequency has a perfect physical sense. Then under wavelet decomposition the two types of frequencies are used to analyze the time-frequency distributions of the biomedical EEG signals with comparison. A set of experiments on publicly available database clearly indicate that the proposed approaches are very promising.