By considering the spectral signature as a sequence, recurrent neural networks (RNNs) have been successfully used to learn discriminative features from hyperspectral images (HSIs) recently. However, most of these models only input the whole spectral bands into RNNs directly, which may not fully explore the specific properties of HSIs. In this paper, we propose a cascaded RNN model using gated recurrent units to explore the redundant and complementary information of HSIs. It mainly consists of two RNN layers. The first RNN layer is used to eliminate redundant information between adjacent spectral bands, while the second RNN layer aims to learn the complementary information from nonadjacent spectral bands. To improve the discriminative ability of the learned features, we design two strategies for the proposed model. Besides, considering the rich spatial information contained in HSIs, we further extend the proposed model to its spectral-spatial counterpart by incorporating some convolutional layers. To test the effectiveness of our proposed models, we conduct experiments on two widely used HSIs. The experimental results show that our proposed models can achieve better results than the compared models.

Facial expression recognition (FER) in the wild received broad concerns in which occlusion and pose variation are two key issues. This paper proposed a global multi-scale and local attention network (MA-Net) for FER in the wild. Specifically, the proposed network consists of three main components: a feature pre-extractor, a multi-scale module, and a local attention module. The feature pre-extractor is utilized to pre-extract middle-level features, the multi-scale module to fuse features with different receptive fields, which reduces the susceptibility of deeper convolution towards occlusion and variant pose, while the local attention module can guide the network to focus on local salient features, which releases the interference of occlusion and non-frontal pose problems on FER in the wild. Extensive experiments demonstrate that the proposed MA-Net achieves the state-of-the-art results on several in-the-wild FER benchmarks: CAER-S, AffectNet-7, AffectNet-8, RAFDB, and SFEW with accuracies of 88.42%, 64.53%, 60.29%, 88.40%, and 59.40% respectively. The codes and training logs are publicly available at https://github.com/zengqunzhao/MA-Net.

Convolutional neural networks (CNNs) have been widely used for hyperspectral image classification. As a common process, small cubes are first cropped from the hyperspectral image and then fed into CNNs to extract spectral and spatial features. It is well known that different spectral bands and spatial positions in the cubes have different discriminative abilities. If fully explored, this prior information will help improve the learning capacity of CNNs. Along this direction, we propose an attention-aided CNN model for spectral–spatial classification of hyperspectral images. Specifically, a spectral attention subnetwork and a spatial attention subnetwork are proposed for spectral and spatial classifications, respectively. Both of them are based on the traditional CNN model and incorporate attention modules to aid networks that focus on more discriminative channels or positions. In the final classification phase, the spectral classification result and the spatial classification result are combined together via an adaptively weighted summation method. To evaluate the effectiveness of the proposed model, we conduct experiments on three standard hyperspectral data sets. The experimental results show that the proposed model can achieve superior performance compared with several state-of-the-art CNN-related models.

We propose a novel spatiotemporal fusion method based on deep convolutional neural networks (CNNs) under the application background of massive remote sensing data. In the training stage, we build two five-layer CNNs to deal with the problems of complicated correspondence and large spatial resolution gaps between MODIS and Landsat images. Specifically, we first learn a nonlinear mapping CNN between MODIS and low-spatial-resolution (LSR) Landsat images and then learn a super-resolution CNN between LSR Landsat and original Landsat images. In the prediction stage, instead of directly taking the outputs of CNNs as the fusion result, we design a fusion model consisting of high-pass modulation and a weighting strategy to make full use of the information in prior images. Specifically, we first map the input MODIS images to transitional images via the learned nonlinear mapping CNN and further improve the transitional images to LSR Landsat images via the fusion model; then, via the learned SR CNN, the LSR Landsat images are supersolved to transitional images, which are further improved to Landsat images via the fusion model. Compared with the previous learning-based fusion methods, mainly referring to the sparse-representation-based methods, our CNNs-based spatiotemporal method has the following advantages: 1) automatically extracting effective image features; 2) learning an end-to-end mapping between MODIS and LSR Landsat images; and 3) generating more favorable fusion results. To examine the performance of the proposed fusion method, we conduct experiments on two representative Landsat-MODIS datasets by comparing with the sparse-representation-based spatiotemporal fusion model. The quantitative evaluations on all possible prediction dates and the comparison of fusion results on one key date in both visual effect and quantitative evaluations demonstrate that the proposed method can generate more accurate fusion results.

This paper proposes a novel deep learning framework named bidirectional-convolutional long short term memory (Bi-CLSTM) network to automatically learn the spectral-spatial features from hyperspectral images (HSIs). In the network, the issue of spectral feature extraction is considered as a sequence learning problem, and a recurrent connection operator across the spectral domain is used to address it. Meanwhile, inspired from the widely used convolutional neural network (CNN), a convolution operator across the spatial domain is incorporated into the network to extract the spatial feature. In addition, to sufficiently capture the spectral information, a bidirectional recurrent connection is proposed. In the classification phase, the learned features are concatenated into a vector and fed to a Softmax classifier via a fully-connected operator. To validate the effectiveness of the proposed Bi-CLSTM framework, we compare it with six state-of-the-art methods, including the popular 3D-CNN model, on three widely used HSIs (i.e., Indian Pines, Pavia University, and Kennedy Space Center). The obtained results show that Bi-CLSTM can improve the classification performance by almost $1.5\%$ as compared to 3D-CNN.