Learning to Balance: Decoupled Siamese Diffusion Transformer for Reference-Based Remote Sensing Image Super-Resolution

TL;DR

This paper introduces DS-DiT, a novel decoupled Siamese diffusion transformer for reference-based remote sensing image super-resolution, effectively balancing reference information utilization and detail recovery.

Contribution

The paper proposes a decoupled attention mechanism and a patch-level weighting module to improve super-resolution quality in remote sensing images.

Findings

DS-DiT outperforms existing methods in quantitative metrics.

The approach enhances visual fidelity of super-resolved images.

The method effectively balances reference reliance and detail recovery.

Abstract

Diffusion-based methods demonstrate significant potential for remote sensing image super-resolution at large scaling factors, particularly in reference-based super-resolution (RefSR) where high-resolution reference images provide critical fine-grained texture priors. However, existing methods often suffer from a trade-off between over-reliance on reference information, which leads to texture artifacts, and underutilization, which results in insufficient detail recovery. To address these issues, we propose DS-DiT, a Decoupled Siamese Diffusion Transformer method that decouples low-resolution and reference interactions at the attention level. By enabling low-resolution structural priors and reference texture information to interact independently with the noisy latent, the framework effectively mitigates inter-source competition. Furthermore, to compensate for the limited local modeling ability of global attention, we introduce a Patch-Level Weights (PLW) module that adaptively modulates the fusion of conditional sources. In addition, this siamese architecture facilitates an autoguidance strategy during inference, which enhances reconstruction by exploiting the prediction discrepancy between strong and weak reference conditions. This approach boosts generation quality without additional training. Experimental results across multiple datasets and scaling factors demonstrate that DS-DiT outperforms existing methods in both quantitative metrics and visual fidelity.