Multi-scale feature enhancement and cross-dimensional attention mechanism fusion network for hyperspectral image classification

BACA-Net: a band-adaptive collaborative attention network for hyperspectral image classification

Semi-supervised deep residual generative adversarial network for hyperspectral image classification

SSSAT-Net: Spectral-Spatial Self-Attention-Based Transformer Network for hyperspectral image classification

Deformation–based dual-branch convolutional fusion network for hyperspectral image classification

Hyperspectral images (HSIs) have very high dimensionality and typically lack sufficient labeled samples, which significantly challenges their processing and analysis. These challenges contribute to the dimensionality curse, making it difficult to describe complex spatial relationships, especially those with non-Euclidean characteristics. This paper presents a multi-scale graph wavelet convolutional network (MS-GWCN) that utilizes a graph wavelet transform within a multi-scale learning framework to accurately capture spatial-spectral features. The MS-GWCN constructs graphs according to 8-neighborhood connectivity schemes, implements spectral graph wavelet transforms for multi-scale decomposition, and aggregates features through multi-scale graph convolutional layers. Our method, the MS-GWCN, demonstrates superior performance compared to existing methodologies. It achieves higher overall accuracy, average accuracy, per-class accuracy, and the Kappa coefficient, as evaluated on three datasets, including the Indian Pines, Salinas, and Pavia University datasets, thereby demonstrating enhanced robustness and generalization capability.

Abstract Although some progress has been made in hyperspectral image (HSI) classification, it still faces many challenges due to limited training samples, insufficient fusion of spectral and spatial information, and consumption of computing resources. In order to effectively address the above problems, this paper proposes a novel combination of dual domain feature extraction and adaptive spectral-spatial feature fusion (DDFE-ASFS), which fully extracts global and local spectral-spatial features and deep high-level semantic features. Firstly, a dual domain feature extraction (DDFE) module is proposed by integrating deep CNNs, fast Fourier transform (FFT) and inverse fast Fourier transform (IFFT), which can fully characterize local and global spectral-spatial and frequency features. Secondly, an efficient adaptive spectral-spatial fusion (EASSF) module is designed to capture the dependency between cross-views by using the attention mechanism while maintaining the consistency of spectral and spatial features. Then, two convolution layers are used to further optimize the features, and pixel-attention and residual path are combined to achieve dynamic fusion of spectral and spatial features. Finally, the spectral graph context optimizer (SGCO) is used to model the long-range dependency relationship, and improve the classification efficiency and accuracy. Extensive evaluations on four popular HSIs show that, with 10\% of the training samples, the proposed method reaches 99.57% average accuracy on the Houston2013 dataset, 99.80% on the Pavia University dataset, 99.85% on the WHU-Hi-HanChuan dataset, and 99.70% on the WHU-Hi-HongHu dataset, superior to some existing advanced technologies.

Hyperspectral image classification (HSIC) involves analyzing high-dimensional data that contain substantial spectral redundancy and spatial noise, which increases the entropy and uncertainty of feature representations. Reducing such redundancy while retaining informative content in spectral–spatial interactions remains a fundamental challenge for building efficient and accurate HSIC models. Traditional deep learning methods often rely on redundant modules or lack sufficient spectral–spatial coupling, limiting their ability to fully exploit the information content of hyperspectral data. To address these challenges, we propose SGFNet, which is a spectral-guided fusion network designed from an information–theoretic perspective to reduce feature redundancy and uncertainty. First, we designed a Spectral-Aware Filtering Module (SAFM) that suppresses noisy spectral components and reduces redundant entropy, encoding the raw pixel-wise spectrum into a compact spectral representation accessible to all encoder blocks. Second, we introduced a Spectral–Spatial Adaptive Fusion (SSAF) module, which strengthens spectral–spatial interactions and enhances the discriminative information in the fused features. Finally, we developed a Spectral Guidance Gated CNN (SGGC), which is a lightweight gated convolutional module that uses spectral guidance to more effectively extract spatial representations while avoiding unnecessary sequence modeling overhead. We conducted extensive experiments on four widely used hyperspectral benchmarks and compared SGFNet with eight state-of-the-art models. The results demonstrate that SGFNet consistently achieves superior performance across multiple metrics. From an information–theoretic perspective, SGFNet implicitly balances redundancy reduction and information preservation, providing an efficient and effective solution for HSIC.

Supervised deep learning methods have been widely utilized in hyperspectral image (HSI) classification tasks. However, acquiring a large number of reliably labeled samples to train deep networks is not always possible in practical HSI applications due to the time-consuming and laborious labeling process. Semi-supervised learning is commonly used in scenarios with insufficient labeled samples. However, semi-supervised models based on a pseudo-label strategy often suffer from error accumulation. To address this issue and improve HSI classification performance with few labeled samples, a semi-supervised deep learning approach is proposed. First, a multi-scale convolutional neural network with accurate discriminative capability is constructed to reduce pseudo-label errors. Then, a new pseudo-label generation strategy based on Dropout is presented, in which feature-level data augmentation is applied by considering multiple predictions of the unlabeled samples to mitigate the error accumulation problem. Finally, the multi-scale CNN and the new pseudo-label strategy are integrated into a unified model to improve HSI classification performance. The experimental results demonstrate that the proposed approach outperforms other semi-supervised methods in the literature on four real HSI datasets with limited labeled samples.

MIAF-Net: Multiscale interactive attention fusion network for hyperspectral image classification

Multi-scale fuzzy self-attention network for hyperspectral image classification with small-samples

CTFN: Multi-scale CNN and transformer with graph encodings fusion network for hyperspectral image classification

Self-Supervised Learning Driven Cross-Domain Feature Fusion Network for Hyperspectral Image Classification

A spectral-spatial attention guided multi-scale convolutional network for hyperspectral image classification

ABSTRACT Deep learning models have obtained great success in hyperspectral image classification tasks. Nevertheless, they are usually vulnerable to adversarial attacks. Some existing works have been made to defend against adversarial attacks in HSI classification. These works primarily focus on lots of adversarial samples and spatial relationships while overlooking the strong long-range dependencies from HSI. To alleviate this problem, we propose a novel spectral spatial mamba adversarial defense network (SSMADNet) for hyperspectral adversarial image classification. It includes a dense involution branch, a spectral mamba branch, and a spatial multiscale mamba branch. The dense involution branch extracts embedding features via three dense involution layers. The spectral mamba branch can learn the spectral sequence information from HSI adversarial samples. The spatial multiscale mamba branch can model the long-range interaction of the whole image. Finally, a spectral spatial feature enhancement module is designed to adaptively enhance useful spectral spatial features of HSI. Extensive experimental results demonstrate that on five HSI adversarial datasets, the proposed SSMADNet achieves higher classification accuracies than state-of-the-art adversarial defense methods. In particular, our method obtains best OA (93.80%) on the Botswana adversarial data, which is much higher than the suboptimal method (OA = 90.30%).

SXSFormer: Spectral Squeeze and Expansion Swin Transformer Network for Hyperspectral Image Classification

Convolutional long short-term memory (ConvLSTM) possesses a remarkable capability of encoding spatial information and capturing long-range dependencies in sequential data. As a result, ConvLSTM has garnered success in hyperspectral image (HSI) classification. Nonetheless, the design of the special gate structures and convolution operations contributes to a high model complexity, making it challenging to deploy in resource-constrained environments. In this article, we propose a fully tensorized ConvLSTM model for HSI spatial-spectral classification under the premise of low complexity. First, we devise a novel and efficient tensor-sequenced convolution in the tensor train (TT) format, called ETTConv. ETTConv can reduce the number of parameters and computations in the standard convolutional layer by tensorizing the convolution kernels and mapping them to a series of smaller ones. Building upon this innovation, we present a novel ETTConvLSTM unit, formed by jointly compressing all weight tensors within the recurrent units. Using it as the fundamental unit, we construct the lightweight a efficient tensor train ConvLSTM 2-D neural network (ETTCL2DNN) model, characterized by reduced complexity without compromised classification performance. Furthermore, to better preserve the joint spatial-spectral structure of HSI data, we extend the ETTConv layer and the ETTConvLSTM unit to their 3-D versions, resulting in a new lightweight a efficient tensor train ConvLSTM 3-D neural network (ETTCL3DNN) model. Extensive quantitative experimental results on three widely used HSI datasets demonstrate the superiority of the proposed methods, exhibiting enhanced classification performance with reduced model complexity.

CenterMamba: Enhancing semantic representation with center-scan Mamba network for hyperspectral image classification

ABSTRACT Hyperspectral images (HSI) contain rich spectral information essential for real-world applications, but traditional methods struggle with limited training data and complexity. Convolutional neural networks (CNNs) also face challenges in capturing global features. This letter proposes a CNN-based model with a global reasoning module (GRM) to integrate local and global features effectively. A spectral–spatial feature extractor (depthwise and pointwise convolutions) captures local details, while a global reasoning layer models long-range relationships. Experiments on four public HSI datasets validate the model’s superior classification performance.

Dual selective fusion transformer network for hyperspectral image classification.