DSXFormer: 双池化光谱挤压-扩展与动态上下文注意力变换器用于超光谱图像分类
DSXFormer: Dual-Pooling Spectral Squeeze-Expansion and Dynamic Context Attention Transformer for Hyperspectral Image Classification
AI总结：
DSXFormer通过双池化光谱挤压-扩展和动态上下文注意力机制，提升超光谱图像分类的光谱可区分性和空间上下文表示能力。
摘要：
超光谱图像分类（HSIC）是一项具有挑战性的任务，由于高光谱维度、复杂的光谱-空间相关性和有限的标记训练样本。尽管基于变换器的模型在HSIC中显示了强大的潜力，但现有方法往往难以在保持计算效率的同时实现充分的光谱可区分性。为了解决这些限制，我们提出了一种新的DSXFormer，即一种新的双池化光谱挤压-扩展变换器，具有动态上下文注意力，用于HSIC。所提出的DSXFormer引入了一个双池化光谱挤压-扩展（DSX）块，该块利用互补的全局平均和最大池化来自适应重新校准光谱特征通道，从而增强光谱可区分性和波段依赖性建模。此外，DSXFormer在基于窗口的变换器架构中结合了动态上下文注意力（DCA）机制，以动态捕捉局部光谱-空间关系，同时显著减少计算开销。光谱双池化挤压-扩展和DCA的联合整合使DSXFormer能够在光谱强调和空间上下文表示之间实现有效的平衡。此外，采用块提取、嵌入和块合并策略来促进高效的多尺度特征学习。在四个广泛使用的超光谱基准数据集上进行了广泛的实验，包括Salinas（SA）、Indian Pines（IP）、Pavia University（PU）和Kennedy Space Center（KSC），结果表明，DSXFormer在分类准确率上始终优于最先进的方法，分别达到99.95%、98.91%、99.85%和98.52%。
英文摘要：
Hyperspectral image classification (HSIC) is a challenging task due to high spectral dimensionality, complex spectral-spatial correlations, and limited labeled training samples. Although transformer-based models have shown strong potential for HSIC, existing approaches often struggle to achieve sufficient spectral discriminability while maintaining computational efficiency. To address these limitations, we propose a novel DSXFormer, a novel dual-pooling spectral squeeze-expansion transformer with Dynamic Context Attention for HSIC. The proposed DSXFormer introduces a Dual-Pooling Spectral Squeeze-Expansion (DSX) block, which exploits complementary global average and max pooling to adaptively recalibrate spectral feature channels, thereby enhancing spectral discriminability and inter-band dependency modeling. In addition, DSXFormer incorporates a Dynamic Context Attention (DCA) mechanism within a window-based transformer architecture to dynamically capture local spectral-spatial relationships while significantly reducing computational overhead. The joint integration of spectral dual-pooling squeeze-expansion and DCA enables DSXFormer to achieve an effective balance between spectral emphasis and spatial contextual representation. Furthermore, patch extraction, embedding, and patch merging strategies are employed to facilitate efficient multi-scale feature learning. Extensive experiments conducted on four widely used hyperspectral benchmark datasets, including Salinas (SA), Indian Pines (IP), Pavia University (PU), and Kennedy Space Center (KSC), demonstrate that DSXFormer consistently outperforms state-of-the-art methods, achieving classification accuracies of 99.95%, 98.91%, 99.85%, and 98.52%, respectively.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
