Convolutional neural networks (CNNs) have demonstrated excellent performance in hyperspectral image (HSI) classification. However, tuning some critical hyper-parameters of a CNN -such as the receptive field (RF) size-presents a major challenge due to the presence of features with different scales in HSIs. Contrary to the conventional design of CNNs, which fizes the RF size, it has been proven that the RF size is modulated by the stimulus and hence depends on the scene being considered. Such a dilemma has been rarely considered in CNN design. In this letter, a new Multi-branch Selective Kernel Network (MSKNet) is introduced, in which the input image is convolved using different RF sizes to create multiple branches, so that the effect of each branch is adjusted by an attention mechanism according to the input contrast. As a result, our newly developed MSKNet is capable of modeling different scales. Our experimental results, conducted on three widelyused HSIs, reveal that the MSKNet can outperform state-ofthe-art CNNs in the context of HSI classification problems. The
Abstract. Convolutional Neural Networks (CNNs) as a well-known deep learning technique has shown a remarkable performance in visual recognition applications. However, using such networks in the area of hyperspectral image classification is a challenging and time-consuming process due to the high dimensionality and the insufficient training samples. In addition, Generative Adversarial Networks (GANs) has attracted a lot of attentions in order to generate virtual training samples. In this paper, we present a new classification framework based on integration of multi-channel CNNs and new architecture for generator and discriminator of GANs to overcome Small Sample Size (SSS) problem in hyperspectral image classification. Further, in order to reduce the computational cost, the methods related to the reduction of subspace dimension were proposed to obtain the dominant feature around the training sample to generate meaningful training samples from the original one. The proposed framework overcomes SSS and overfitting problem in classifying hyperspectral images. Based on the experimental results on real and well-known hyperspectral benchmark images, our proposed strategy improves the performance compared to standard CNNs and conventional data augmentation strategy. The overall classification accuracy in Pavia University and Indian Pines datasets was 99.8% and 94.9%, respectively.
Generative Adversarial Networks (GANs) have shown striking performances in computer vision applications to augment Virtual Training Samples (VTS). However, the VTS generating by GANs in the context of hyperspectral image classification suffer from structural inconsistency due to the insufficient number of training samples in order to learn high-order features from discriminator. This work addresses the scarcity of training samples by designing a GAN, in which the performance of discriminator is improved to produce more structurally coherent VTS. In the proposed method, by splitting the discriminator into two parts, GAN undertakes two tasks: the main task is to learn distinguish between real and fake samples and the auxiliary task is to learn distinguish structurally corrupted and real samples. With this setup, GAN will produce real-like VTS with a higher variation than conventional GAN. Further, in order to reduce the computational cost, subspace based dimension reduction was performed to obtain the dominant features around the training samples to generate meaningful patterns from the original ones to be used in the learning phase. Based on the experimental results on real and well-known hyperspectral benchmark images, the proposed method improves the performance compared with GANs-related and conventional data augmentation strategies.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.