Region growing based joint sparsity model for hyperspectral image analysis

Owing to the high spectral sampling, the spectral information in hyperspectral imagery (HSI) is often highly correlated and contains redundancy. Motivated by the recent success of sparsity preserving based dimensionality reduction (DR) techniques in both computer vision and remote sensing image analysis community, a novel supervised nonparametric sparse discriminant analysis (NSDA) algorithm is presented for HSI classification. The objective function of NSDA aims at preserving the within-class sparse reconstructive relationship for within-class compactness characterization and maximizing the nonparametric between-class scatter simultaneously to enhance discriminative ability of the features in the projected space. Essentially, it seeks for the optimal projection matrix to identify the underlying discriminative manifold structure of a multiclass dataset. Experimental results on one visualization dataset and three recorded HSI dataset demonstrate NSDA outperforms several state-of-the-art feature extraction methods for HSI classification.

Sparse representation in compressed sensing is a hot topic in signal processing and artificial intelligence due to its success in various applications. A general classification algorithm based on sparse representation theory named Sparse Representation Classification (SRC) was successfully applied in face recognition. In this paper, we research the issue of facial expression recognition (FER) via SRC. Extensive experiments of FER via SRC algorithm are carried out to study the performance of sparse representation theory for FER. The comparison of SRC algorithm with various traditional algorithms such as two-dimensional PCA as well as curvelet transform for FER is also given. Support vector machine is used for expression classification. The paper also studies the robustness of SRC algorithm for FER to noises. Satisfactory results are obtained for FER via sparse representation. The experiment results show the effectiveness of SRC algorithm on FER.

Hyperspectral imagery (HSI) denoising is an important preprocessing step for real-world applications. Recently, sparse representation and low-rank representation based methods are proven effective in HSI denoising. However, most of these approaches only consider the low-rankness in the spectral domain and the sparsity in coding matrix. They have ignored the property that the coding matrix of each atom is also low-rank, i.e., low-rankness also exists in the spatial domain. In this paper, a reweighed sparse low-rank nonnegative tensor factorization (RSLRNTF) method is proposed to restore an HSI. It takes an HSI as a third-order tensor and factorizes it into the combination of a few component tensors where each one is the outer product of a low-rank matrix (coding matrix) and a vector (atom). Additionally, a reweighed L1 norm is added to coding matrices to enforce their sparsity. The low-rankness in both the spatial domain and the spectral domain as well as sparsity in the spatial domain improve the denoising performance. Furthermore, the nonnegativity in both coding matrices and dictionary leads to parts-based representation of HSI, which facilitates preserving local fine structure information. Experimental results on synthetic data and real-world data demonstrate the superiority of proposed method.

Pan-sharpening - fusing the spatial and spectral information between panchromatic (PAN) and multispectral (MSI) or hyperspectral (HSI) imagery of a common scene is a hot topic in remote sensing due to a wide range of applications such as target detection, vegetation monitoring, and subsurface detection (e.g. landmines), among others. However, the focus of panchromatic sharpening is generally placed on visual quality of the resulting image and image-wide summary spectral accuracy metrics. Here we are interested in radiometrically accurate panchromatic sharpening of hyperspectral imagery with particular emphasis on spectral algorithm performance. Four pansharpening algorithms are applied to hyperspectral imagery and evaluated for spectral/radiometric fidelity. Two datasets from SHARE2012 were used: one which features rural scene elements and one which features an urban scene. Target detection was also performed to evaluate algorithm sharpening performance. We find that although visually the performance of the four algorithms were roughly similar, they differ in spectral/radiometric fidelity as well as performance in ACE target detection.

Based on multiple maximum scatter difference discrimination Dictionary learning, a novel face recognition algorithm is proposed. Dictionary used for sparse coding plays a key role in sparse representation classification. In this paper, a multiple maximum scatter difference discriminated criterion is used for dictionary learning. During the process of dictionary learning, the multiple maximum scatter difference computes its discriminated vectors from both the range of the between class scatter matrix and the null space of the within-class scatter matrix. The proposed algorithm is theoretically elegant and easy to calculate. Extensive experimental studies conducted on the AR database and Extended Yale Database B in comparison with existing basic sparse representation and other classification methods, it shows that the performance is a little better than the original sparse representation methods with lower complexity.

Semi-supervised two phase test sample sparse representation classifier

Estimation of corn yield based on hyperspectral imagery and convolutional neural network

For the instrument limitation and imperfect imaging optics, it is difficult to acquire high spatial resolution hyperspectral imagery. Low spatial resolution will result in a lot of mixed pixels and greatly degrade the detection and recognition performance, affect the related application in civil and military fields. As a powerful statistical image modeling technique, sparse representation can be utilized to analyze the hyperspectral image efficiently. Hyperspectral imagery is intrinsically sparse in spatial and spectral domains, and image super-resolution quality largely depends on whether the prior knowledge is utilized properly. In this article, we propose a novel hyperspectral imagery super-resolution method by utilizing the sparse representation and spectral mixing model. Based on the sparse representation model and hyperspectral image acquisition process model, small patches of hyperspectral observations from different wavelengths can be represented as weighted linear combinations of a small number of atoms in pre-trained dictionary. Then super-resolution is treated as a least squares problem with sparse constraints. To maintain the spectral consistency, we further introduce an adaptive regularization terms into the sparse representation framework by combining the linear spectrum mixing model. Extensive experiments validate that the proposed method achieves much better results.

Recently, deep learning-based methods have drawn increasing attention in hyperspectral imagery (HSI) classification, due to their strong nonlinear mapping capability. However, these methods suffer from a time-consuming training process because of many network parameters. In this paper, the concept of broad learning is introduced into HSI classification. Firstly, to make full use of abundant spectral and spatial information of hyperspectral imagery, hierarchical guidance filtering is performing on the original HSI to get its spectral-spatial representation. Then, the class-probability structure is incorporated into the broad learning model to obtain a semi-supervised broad learning version, so that limited labeled samples and many unlabeled samples can be utilized simultaneously. Finally, the connecting weights of broad structure can be easily computed through the ridge regression approximation. Experimental results on three popular hyperspectral imagery datasets demonstrate that the proposed method can achieve better performance than deep learning-based methods and conventional classifiers.

The hyperspectral/spatial detection of edges (HySPADE) algorithm, originally published in 2004 [1], has been modified and applied to a wider diversity of hyperspectral imagery (HSI) data. As originally described in [1], HySPADE operates by converting the naturally two-dimensional edge detection process based on traditional image analysis methods into a series of one-dimensional edge detections based on spectral angle. The HySPADE algorithm: i) utilizes spectral signature information to identify edges; ii) requires only the spectral information of the HSI scene data and does not require a spectral library nor spectral matching against a library; iii) facilitates simultaneous use of all spectral information; iv) does not require endmember or training data selection; v) generates multiple, independent data points for statistical analysis of detected edges; vi) is robust in the presence of noise; and vii) may be applied to radiance, reflectance, and emissivity data--though it is applied to radiance and reflectance spectra (and their principal components transformation) in this report. HySPADE has recently been modified to use Euclidean distance values as an alternative to spectral angle. It has also been modified to use an N x N-pixel sliding window in contrast to the 2004 version which operated only on spatial subset image chips. HySPADE results are compared to those obtained using traditional (Roberts and Sobel) edge-detection methods. Spectral angle and Euclidean distance HySPADE results are superior to those obtained using the traditional edge detection methods; the best results are obtained by applying HySPADE to the first few, information-containing bands of principal components transformed data (both radiance and reflectance). However, in practice, both the Euclidean distance and spectral angle versions of HySPADE should be applied and their results compared. HySPADE results are shown; extensions of the HySPADE concept are discussed as are applications for HySPADE in HSI analysis and exploitation.

Facial expressions play important role in human communication. The understanding of facial expression is a basic requirement in the development of next generation human computer interaction systems. Researches show that the intrinsic facial features always hide in low dimensional facial subspaces. This paper presents facial parts based facial expression recognition system with sparse representation classifier. Sparse representation classifier exploits sparse representation to select face features and classify facial expressions. The sparse solution is obtained by solving l¹ -norm minimization problem with constraint of linear combination equation. Experimental results show that sparse representation is efficient for facial expression recognition and sparse representation classifier obtain much higher recognition accuracies than other compared methods.

Multivariate classification techniques have been widely applied to decode brain states using functional magnetic resonance imaging (fMRI). Due to variabilities in fMRI data and the limitation of the collection of human fMRI data, it is not easy to train an efficient and robust supervised-learning classifier for fMRI data. Among various classification techniques, sparse representation classifier (SRC) exhibits a state-of-the-art classification performance in image classification. However, SRC has rarely been applied to fMRI-based decoding. This study aimed to improve SRC using unlabeled testing samples to allow it to be effectively applied to fMRI-based decoding. We proposed a semisupervised-learning SRC with an average coefficient (semiSRC-AVE) method that performed the classification using the average coefficient of each class instead of the reconstruction error and selectively updated the training dataset using new testing data with high confidence to improve the performance of SRC. Simulated and real fMRI experiments were performed to investigate the feasibility and robustness of semiSRC-AVE. The results of the simulated and real fMRI experiments showed that semiSRC-AVE significantly outperformed supervised learning SRC with an average coefficient (SRC-AVE) method and showed better performance than the other three semisupervised learning methods.

Owing to complicated manual design and feature extraction processes of mechanical multi-fault classification and recognition methods, single-class classifiers do not adapt well to the high dimensionality and high self-similarity among mechanical vibration signals with nonlinear characteristics, resulting in insufficient classification accuracy. Combining the adaptive learning and feature extraction capabilities of the wavelet kernel-based extreme learning machine (WKELM) with the sparse representation classifier (SRC), this paper proposes a hybrid artificial bee colony (ABC)-optimized Morlet wavelet kernel-based extreme learning machine (ABC-MWKELM) with SRC. The two-level classifier is capable of cooperatively classifying multiple mechanical faults. First, a Morlet wavelet kernel-based extreme learning machine (MWKELM) model that matches the kernel function construction condition is presented. An ABC optimization algorithm is adopted to optimize critical parameters of the MWKELM model, thereby improving the accuracy and generalization ability of the model. In addition, the ABC algorithm is used to optimize the MWKELM model as a pre-classifier for classifying test samples. According to the posterior probability of each training sample category output, the threshold decision method selects candidate categories with a high confidence coefficient. When the preset meets the threshold, the classification result of the ABC-optimized MWKELM is directly adopted. Otherwise, the candidate training samples are used to construct a dictionary and the SRC is used to classify the test samples further. The experimental results demonstrate that compared with the ELM, WKELM, (ELM-SRC), and radial basis function (RBF)-KELM and SRC (RBFKELM-SRC), the method presented in this paper has significant advantages in terms of accuracy of mechanical multi-fault classification. The validity and applicability of this method in mechanical multi-fault classification were verified.

Noise reduction is always an active research area in image processing due to its importance for the sequential tasks such as object classification and detection. In this paper, we develop a sparse representation based noise reduction method for hyperspectral imagery, which is dependent on the assumption that the non-noise component in the signal can be approximated by only a small number of atoms in a dictionary while noise component has not this property. The main contribution of the paper is in introducing nonlocal similarity and spectral-spatial structure of hyperspectral imagery into sparse representation. Non-locality means the self-similarity of image, by which the whole image can be partitioned into some groups containing similar patches. The similar patches in each group is sparsely represented with shared atoms making the signal and noise more easily separated. Sparse representation with spectral-spatial structure can exploit spectral and spatial joint correlations of hyperspectral imagery also making the signal and noise more distinguished, in which 3-D blocks are instead of 2-D patches for sparse coding. The experimental results indicate that the proposed method has a good quality of restoring the true signal from the noisy observation.

In order to improve the classification performance for hyperspectral image (HSI), a sparse representation classifier based on spectral information divergence (SID) is proposed. SID measures the discrepancy of probabilistic behaviors between the spectral signatures of two pixels from the aspect of information theory, which can be more effective in preserving spectral properties. Thus, the new method measures the similarity between the reconstructed pixel and the true pixel by SID instead of by the L2 norm used in traditional sparse model. Moreover, the spatial coherency across neighboring pixels sharing a common sparsity pattern is taken into account during the construction of SID-based joint sparse representation model. We propose a new version of the orthogonal matching pursuit method to solve SID-based recovery problems. The proposed SID-based algorithms are applied to real HSI for classification. Experimental results show that our algorithms outperform the classical sparse representation based classification algorithms in most cases.