18 September 2024 in IEEE Transactions on Geoscience and Remote Sensing
In the hyperspectral image classification (HSIC) task, the most commonly used model validation paradigm is partitioning the training-test dataset through pixel-wise random sampling. By training on a small amount of data, the deep learning model can achieve almost perfect accuracy. However, in our experiments, we found that the high accuracy was reached because the training and test datasets share a lot of information. On non-overlapping dataset partitions, well-performing models suffer significant performance degradation. To this end, we propose a spectral-spatial axial aggregation transformer model, namely SaaFormer, that preserves generalization across dataset partitions. SaaFormer applies a multi-level spectral extraction structure to segment the spectrum into multiple spectrum clips, such that the wavelength continuity of the spectrum across the channel are preserved. For each spectrum clip, the axial aggregation attention mechanism, which integrates spatial features along multiple spectral axes is applied to mine the spectral characteristic. The multi-level spectral extraction and the axial aggregation attention emphasize spectral characteristic to improve the model generalization. The experimental results on five publicly available datasets demonstrate that our model exhibits comparable performance on the random partition, while significantly outperforming other methods on non-overlapping partitions. Moreover, SaaFormer shows excellent performance on background classification.
2 March 2026 in SSRN Electronic Journal
Remote sensing image Pansharpening is a crucial technique for generating high-resolution multispectral images, widely applied in fields such as land classification and disaster monitoring. However, existing deep pansharpening networks exhibit severe vulnerability to subtle adversarial perturbations, which leads to a steep decline in the quality of fusion results in practical applications. To address this inherent flaw, this paper proposes a dictionary mapping-based pansharpening paradigm and constructs DGSPNet. We designed the Dictionary-Guided Cross-Projection Module (DGCP) to stabilize intermediate features by constraining the network’s intermediate feature representations within a structured subspace via learnable dictionary projection, and integrated the Frequency Decomposition Attention Module (FDAM) to enhance spatial details while preserving spectral consistency. A Response Consistency Loss was introduced during the training phase for adversarial training. Experimental results demonstrate that the proposed model exhibits superior performance and significant robustness in both clean and adversarial test scenarios, achieving a simultaneous improvement in both robustness and fusion quality compared to existing methods.