Convolutional neural networks have garnered increasing interest for the supervised classification of hyperspectral imagery. However, images with a wide variety ofspatial land-cover sizes can hinder the feature-extraction ability of traditional convolutional networks. Consequently, many approaches intended to extract multiscale features have emerged; these techniques typically extract features in multiple parallel branches using convolutions of differing kernel sizes with concatenation or addition employed to fuse the features resulting from the various branches. In contrast, the present work explores a multiscale spatial-spectral feature-extraction network that operates in a more granular manner. Specifically, in the proposed network, a multibranch structure expands the convolutional receptive fields through the partitioning of input feature maps, applying hierarchical connections across the partitions, crosschannel feature fusion via pointwise convolution, and depthwise three-dimensional (3-D) convolutions for feature extraction. Experimental results reveal that the proposed multiscale spatial-spectral feature-fusion network outperforms other state-of-the-art networks at the supervised classification of hyperspectral imagery while being robust to limited training data.

/pdf/multiscale-spatial-spectral-feature-extraction-network-for-29ymct82.pdf

Multiscale Spatial-Spectral Feature Extraction Network for Hyperspectral Image Classification

Infrared and visible image fusion using a generative adversarial network with a dual-branch generator and matched dense blocks

In recent years, convolutional neural networks (CNNs) have been widely used in hyperspectral image classification and have achieved good performance. However, the high dimensions and few samples of hyperspectral remote sensing images tend to be the main factors restricting improvements in classification performance. At present, most advanced classification methods are based on the joint extraction of spatial and spectral features. In this article, an improved dense block based on a multiscale spectral pyramid (MSSP) is proposed. This method uses the idea of multiscale and group convolution of the convolution kernel, which can fully extract spectral information from hyperspectral images. The designed MSSP is the main unit of the spectral dense block (called MSSP Block). Additionally, a short connection with nonlinear transformation is introduced to enhance the representation ability of the model. To demonstrate the effectiveness of the proposed dual-branch multiscale spectral attention network, some experiments are conducted on five commonly used datasets. The experimental results show that, compared with some state-of-the-art methods, the proposed method can provide better classification performance and has strong generalization ability.

/pdf/hyperspectral-image-classification-based-on-dual-branch-fez0hitk90.pdf

Hyperspectral Image Classification Based on Dual-Branch Spectral Multiscale Attention Network

Deep Neural Network (DNN) models have brought significant performance gains to machine learning tasks. Nevertheless, the huge storage and computational costs of high-performance models severely limit their deployment on resource-limited embedded devices. Knowledge distillation (KD) is the mainstream approach for models compression, but the existing KD framework has some limitations. First, the student model underperforms when there is a considerable capacity gap between the student and teacher models, which reduces the model compression rate in knowledge distillation. Second, compared with training from scratch, the billions of additional forward propagation of the complex teacher model is time-consuming, resulting longer train time in the distillation process. This paper proposes AKD, an efficient difference-adaptive knowledge distillation framework. AKD consists of two modules. The distillation module uses a modified auxiliary teaching architecture as the backbone, and the adaptive module introduces a difference-adaptive method, which can adjust the teacher model adaptively according to the continuously improving representation ability of the student model. Experimental results on different kinds of datasets and models indicate that AKD not only improves the efficiency of distillation, but also enhances the performance of the compact student model and raise the compression rate.

Yi Xiong

Papers

Hyperspectral Image Classification Based on Dual-Branch Spectral Multiscale Attention Network

Difference-Adaptive Knowledge Distillation for Resource-Constrained Embedded Devices