ALBERT

Description: In recent years, endmember variability has received much attention in the field of hyperspectral unmixing. To solve the problem caused by the inaccuracy of the endmember signature, the endmembers are usually modeled to assume followed by a statistical distribution. However, those distribution-based methods only use the spectral information alone and do not fully exploit the possible local spatial correlation. When the pixels lie on the inhomogeneous region, the abundances of the neighboring pixels will not share the same prior constraints. Thus, in this paper, to achieve better abundance estimation performance, a method based on the Gaussian mixture model (GMM) and spatial group sparsity constraint is proposed. To fully exploit the group structure, we take the superpixel segmentation (SS) as preprocessing to generate the spatial groups. Then, we use GMM to model the endmember distribution, incorporating the spatial group sparsity as a mixed-norm regularization into the objective function. Finally, under the Bayesian framework, the conditional density function leads to a standard maximum a posteriori (MAP) problem, which can be solved using generalized expectation-maximization (GEM). Experiments on simulated and real hyperspectral data demonstrate that the proposed algorithm has higher unmixing precision compared with other state-of-the-art methods.

Description: Many deep learning models, such as convolutional neural network (CNN) and recurrent neural network (RNN), have been successfully applied to extracting deep features for hyperspectral tasks. Hyperspectral image classification allows distinguishing the characterization of land covers by utilizing their abundant information. Motivated by the attention mechanism of the human visual system, in this study, we propose a spectral-spatial attention network for hyperspectral image classification. In our method, RNN with attention can learn inner spectral correlations within a continuous spectrum, while CNN with attention is designed to focus on saliency features and spatial relevance between neighboring pixels in the spatial dimension. Experimental results demonstrate that our method can fully utilize the spectral and spatial information to obtain competitive performance.

Description: With the rapid developments of hyperspectral imaging, the cost of collecting hyperspectral data has been lower, while the demand for reliable and detailed hyperspectral annotations has been much more substantial. However, limited by the difficulties of labelling annotations, most existing hyperspectral image (HSI) classification methods are trained and evaluated on a single hyperspectral data cube. It brings two significant challenges. On the one hand, many algorithms have reached a nearly perfect classification accuracy, but their trained models are hard to generalize to other datasets. On the other hand, since different hyperspectral datasets are usually not collected in the same scene, different datasets will contain different classes. To address these issues, in this paper, we propose a new paradigm for HSI classification, which is training and evaluating separately across different hyperspectral datasets. It is of great help to labelling hyperspectral data. However, it has rarely been studied in the hyperspectral community. In this work, we utilize a three-phase scheme, including feature embedding, feature mapping, and label reasoning. More specifically, we select a pair of datasets acquired by the same hyperspectral sensor, and the classifier learns from one dataset and then evaluated it on the other. Inspired by the latest advances in zero-shot learning, we introduce label semantic representation to establish associations between seen categories in the training set and unseen categories in the testing set. Extensive experiments on two pairs of datasets with different comparative methods have shown the effectiveness and potential of zero-shot learning in HSI classification.