Spectral Dynamic Attention Network for Hyperspectral Image Super-Resolution

arXiv cs.CV / 5/1/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes Spectral Dynamic Attention Network (SDANet) to improve hyperspectral image super-resolution by reducing spectral redundancy and increasing non-linear modeling capacity.
SDANet uses a Dynamic Channel Sparse Attention (DCSA) module that adaptively computes channel-wise correlations and applies data-dependent sparsification to keep only the most informative interactions.
It also introduces a Frequency-Enhanced Feed-Forward Network (FE-FFN) that jointly leverages spatial and frequency-domain features to improve expressiveness.
Experiments on two benchmark datasets show SDANet achieves state-of-the-art hyperspectral image super-resolution performance while remaining competitively efficient.
The authors plan to publicly release the code at the provided GitHub repository to support reproducibility and further research.

Abstract

Hyperspectral image super-resolution is essential for enhancing the spatial fidelity of HSI data, yet existing deep learning methods often struggle with substantial spectral redundancy and the limited non-linear modeling capacity of standard feed-forward networks (FFNs). To address these challenges, we propose Spectral Dynamic Attention Network (SDANet), a framework designed to adaptively suppress redundant spectral interactions. SDANet integrates two key components: 1) Dynamic Channel Sparse Attention (DCSA) module that computes channel-wise correlations and selectively preserves the most informative attention responses through dynamic and data-dependent sparsification. 2) Frequency-Enhanced Feed-Forward Network (FE-FFN) that jointly models spatial and frequency-domain representations to enhance non-linear expressiveness. Extensive experiments on two benchmark datasets demonstrate that SDANet achieves state-of-the-art HISR performance while maintaining competitive efficiency. The code will be made publicly available at https://github.com/oucailab/SDANet.