Skip to content
Open access

Residual Learning and Feature Weighing-Based Mamba Network for Hyperspectral Image Super Resolution

2026 · IEEE Geoscience and Remote Sensing Letters · Vol 23, pp. 5508205-5508205 · 0 citations · 17 references

Abstract

The hyperspectral image super-resolution (HSI-SR) enhances spatial resolution while maintaining the information-rich spectral bands in HSI. However, managing the computational overload of spectral bands and capturing the long-range spatial dependencies remains a challenging task. To address these issues, this letter proposed a residual learning and feature weighing-based Mamba network (RFWMN) for efficient HSI-SR. The RFWMN integrates residual learning to accelerate training convergence and a feature weighing approach to emphasize significant features, thereby reducing computational complexity. Furthermore, the RFWMN utilized the 2-D selective scan (SS2D)-based Mamba state-space model (SSM) with multiple receptive fields to capture the long-range spatial dependencies with linear computational complexity. Besides, a triple attention block (TAB) is designed by integrating spatial, spectral, and pixelwise attention to further improve reconstruction performance. The proposed RFWMN outperforms other models on three publicly available datasets across metrics, such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), spectral angle mapper (SAM), Erreur Relative Globale Adimensionnelle de Synthése (ERAGS), and universal image quality index (UIQI), while maintaining comparatively lower parameter counts, floating-point operations per second (FLOPs), and graphics processing unit (GPU) memory consumption.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.