Answers are grounded in this paper and cite their sources.
Multi-modal remote sensing image matching is a crucial task with broad application potential. However, substantial nonlinear radiometric differences between multi-modal images pose significant challenges, often leading to mismatches. To tackle these challenges, this paper introduces WinMRSI, a window attention-based multi-modal remote sensing image matching method designed to enhance cross-modal feature extraction and information interaction. For feature extraction, a siamese network with Discrete Cosine Transform (DCT) is employed to model inter-channel dependencies and extract multi-scale features from cross-modal images. Additionally, a dual-branch network is designed to capture contextual dependencies while refining local feature representations. For information interaction, WinMRSI integrates a window attention mechanism to strengthen fine-grained feature fusion within highly relevant windows, enabling the model to focus on discriminative regions. Furthermore, a multi-level matching module progressively refines matching accuracy in a coarse-to-fine manner across window, patch, and pixel levels. Extensive evaluations on benchmark datasets demonstrate that WinMRSI achieves state-of-the-art performance in multi-modal remote sensing image matching. Ablation studies further validate the effectiveness of each component in WinMRSI.
No takes yet. Share an insight, caveat, or question.
Di et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: