This research proposes an adaptive state space network to improve image quality in resource-constrained settings, highlighting its effectiveness across diverse degradation types.
Key Points
The DWMamba network enhances image quality by addressing diverse degradation types through an efficient architecture, improving detail restoration.
Key experiments show that DWMamba outperforms existing methods in both qualitative and quantitative measures, demonstrating adaptability under various lighting conditions.
An Adaptive State Space Module is implemented, featuring dual-stream monitoring and soft fusion to effectively capture global dependencies in image processing.
The Structure-guided Residual Fusion module aids in selectively fusing features, significantly enhancing low-light texture recovery and detail restoration.