Deep learning methods have recently been successfully explored for hyperspectral image (HSI) classification. However, training a deep-learning classifier notoriously requires hundreds or thousands of labeled samples. In this paper, a deep few-shot learning method is proposed to address the small sample size problem of HSI classification. There are three novel strategies in the proposed algorithm. First, spectral–spatial features are extracted to reduce the labeling uncertainty via a deep residual 3-D convolutional neural network. Second, the network is trained by episodes to learn a metric space where samples from the same class are close and those from different classes are far. Finally, the testing samples are classified by a nearest neighbor classifier in the learned metric space. The key idea is that the designed network learns a metric space from the training data set. Furthermore, such metric space could generalize to the classes of the testing data set. Note that the classes of the testing data set are not seen in the training data set. Four widely used HSI data sets were used to assess the performance of the proposed algorithm. The experimental results indicate that the proposed method can achieve better classification accuracy than the conventional semisupervised methods with only a few labeled samples.

Deep Few-Shot Learning for Hyperspectral Image Classification

Classification of hyperspectral image (HSI) is an important research topic in the remote sensing community. Significant efforts (e.g., deep learning) have been concentrated on this task. However, it is still an open issue to classify the high-dimensional HSI with a limited number of training samples. In this paper, we propose a semi-supervised HSI classification method inspired by the generative adversarial networks (GANs). Unlike the supervised methods, the proposed HSI classification method is semi-supervised, which can make full use of the limited labeled samples as well as the sufficient unlabeled samples. Core ideas of the proposed method are twofold. First, the three-dimensional bilateral filter (3DBF) is adopted to extract the spectral-spatial features by naturally treating the HSI as a volumetric dataset. The spatial information is integrated into the extracted features by 3DBF, which is propitious to the subsequent classification step. Second, GANs are trained on the spectral-spatial features for semi-supervised learning. A GAN contains two neural networks (i.e., generator and discriminator) trained in opposition to one another. The semi-supervised learning is achieved by adding samples from the generator to the features and increasing the dimension of the classifier output. Experimental results obtained on three benchmark HSI datasets have confirmed the effectiveness of the proposed method , especially with a limited number of labeled samples.

https://www.mdpi.com/2072-4292/9/10/1042/pdf

Generative Adversarial Networks-Based Semi-Supervised Learning for Hyperspectral Image Classification

Most traditional approaches classify hyperspectral image (HSI) pixels relying only on the spectral values of the input channels. However, the spatial context around a pixel is also very important and can enhance the classification performance. In order to effectively exploit and fuse both the spatial context and spectral structure, we propose a novel two-stream deep architecture for HSI classification. The proposed method consists of a two-stream architecture and a novel fusion scheme. In the two-stream architecture, one stream employs the stacked denoising autoencoder to encode the spectral values of each input pixel, and the other stream takes as input the corresponding image patch and deep convolutional neural networks are employed to process the image patch. In the fusion scheme, the prediction probabilities from two streams are fused by adaptive class-specific weights, which can be obtained by a fully connected layer. Finally, a weight regularizer is added to the loss function to alleviate the overfitting of the class-specific fusion weights. Experimental results on real HSIs demonstrate that the proposed two-stream deep architecture can achieve competitive performance compared with the state-of-the-art methods.

Two-Stream Deep Architecture for Hyperspectral Image Classification

Semisupervised learning is widely used in hyperspectral image classification to deal with the limited training samples, however, some more information of hyperspectral image should be further explored. In this paper, a novel semisupervised classification based on multi-decision labeling and deep feature learning is presented to exploit and utilize as much information as possible to realize the classification task. First, the proposed method takes two decisions to pre-label each unlabeled sample: local decision based on weighted neighborhood information is made by the surrounding samples, and global decision based on deep learning is performed by the most similar training samples. Then, some unlabeled ones with high confidence are selected to extent the training set. Finally, self decision, which depends on the self features exploited by deep learning, is employed on the updated training set to extract spectral-spatial features and produce classification map. Experimental results with real data indicate that it is an effective and promising semisupervised classification method for hyperspectral image.

Semisupervised classification for hyperspectral image based on multi-decision labeling and deep feature learning

Deep learning has achieved great success in hyperspectral image classification. However, when processing new hyperspectral images, the existing deep learning models must be retrained from scratch with sufficient samples, which is inefficient and undesirable in practical tasks. This paper aims to explore how to accurately classify new hyperspectral images with only a few labeled samples, i.e., the hyperspectral images few-shot classification. Specifically, we design a new deep classification model based on relational network and train it with the idea of meta-learning. Firstly, the feature learning module and the relation learning module of the model can make full use of the spatial–spectral information in hyperspectral images and carry out relation learning by comparing the similarity between samples. Secondly, the task-based learning strategy can enable the model to continuously enhance its ability to learn how to learn with a large number of tasks randomly generated from different data sets. Benefitting from the above two points, the proposed method has excellent generalization ability and can obtain satisfactory classification results with only a few labeled samples. In order to verify the performance of the proposed method, experiments were carried out on three public data sets. The results indicate that the proposed method can achieve better classification results than the traditional semisupervised support vector machine and semisupervised deep learning models.

https://www.mdpi.com/2072-4292/12/6/923/pdf

Deep Relation Network for Hyperspectral Image Few-Shot Classification

Graph-based classification algorithms have gained increasing attention in semi-supervised classification. Nevertheless, the graph cannot fully represent the inherent spatial distribution of the data. In this paper, a new classification methodology based on the spatial-spectral Label Propagation is proposed for semi-supervised classification of hyperspectral imagery. The spatial information was used in two aspects: on the one hand, the spatial features extracted by a 2-D Gabor filter were stacked with spectral features; on the other hand, the width of the Gaussian function, which was used to construct graph, was determined with an adaptive method. Subsequently, the unlabeled samples from the spatial neighbors of the labeled samples were selected and the spatial graph was constructed based on spatial smoothness. Finally, labels were propagated from labeled samples to unlabeled samples with spatial-spectral graph to update the training set for a basic classifier (e.g., Support Vector Machine, SVM). Experiments on four hyperspectral datasets show that the proposed Spatial-Spectral Label Propagation based on the SVM (SS-LPSVM) can effectively represent the spatial information in the framework of semi-supervised learning and consistently produces greater classification accuracy than the standard SVM, the Laplacian Support Vector Machine (LapSVM), Transductive Support Vector Machine (TSVM) and the Spatial-Contextual Semi-Supervised Support Vector Machine (SCS 3 VM).

Semi-supervised classification for hyperspectral imagery based on spatial-spectral Label Propagation

Spectral unmixing has been an important technique for hyperspectral imagery processing. In traditional spectral unmixing methods that are based on the linear spectral mixture model (LSMM), unmixing accuracy is limited by the inherent deficiency of the model. It was shown that the support vector machine (SVM) can be extended for spectral unmixing, based on the advantage that the SVM model can accommodate the variations within a relative pure class by using multiple pure samples instead of a single endmember for one class. In the SVM model, class label errors are considered in constraints. However, the errors concerned in spectral unmixing are the unmixing residue instead of the class label ones. This letter presents a method of imposing unmixing residue constraints on the least squares SVM unmixing model. The related problems, including deducing the closed-form solution and substituting the single endmember for multiple ones, were studied together. Experiments showed that the new SVM model was superior to the original SVM as well as the traditional LSMM in terms of unmixing residue, fractional abundance, and confused matrix criterions.

Spectral Unmixing Model Based on Least Squares Support Vector Machine With Unmixing Residue Constraints

Semisupervised learning has shown its great potential in land cover mapping. It exploits the information of unlabeled training samples and converts those samples to labeled training samples to enhance classification. In this paper, the spatial information extracted by a two-dimensional (2-D) Gabor filter was stacked with spectral information first, and then the spatial neighborhood information of labeled training samples was combined with active learning (AL) algorithm to select the most useful and informative samples, which were used as the unlabeled set to aid the probability model-based supervised support vector machine (SVM). Experiments on two hyperspectral datasets showed that the spatial-spectral information-based semisupervised classification algorithm ($\bf{S}^{\bf{2}} \bf{ISC}$) can produce high classification accuracy with a small number of labeled samples, and outperformed the compared semisupervised algorithms.

Spatial–Spectral Information-Based Semisupervised Classification Algorithm for Hyperspectral Imagery

Iterative Error Analysis (IEA) widely known as a good endmember extraction (EE) algorithm. It is robust, automatic and free of data transformation. However, IEA is faced with risks in some cases due to the sole use of unmxing distance, and its speed is lowed down by the iteration-based linear spectral mixture analysis (LSMA). To make IEA algorithm faster and more robust, its modified version is proposed based on two substitutions. One is substituting integrated distance for unmxing distance, which makes the algorithm more robust. The other is substituting SVM-based multiple endmember spectral mixture analysis (MESMA) for iteration-based LSMA, which speeds up the algorithm greatly. Experiments show that the modified IEA algorithm outperforms original one in terms of both robustness and running speed.

Ying Wang

Papers

Semi-supervised classification for hyperspectral imagery based on spatial-spectral Label Propagation

Spectral Unmixing Model Based on Least Squares Support Vector Machine With Unmixing Residue Constraints

Spatial–Spectral Information-Based Semisupervised Classification Algorithm for Hyperspectral Imagery

Endmember extraction based on modified Iterative Error Analysis