SDCoNet: Saliency-Driven Multi-Task Collaborative Network for Remote Sensing Object Detection

TL;DR

SDCoNet is a novel multi-task network that enhances remote sensing object detection by integrating saliency-driven super-resolution with collaborative feature sharing, improving accuracy on small objects in low-quality images.

Contribution

The paper introduces SDCoNet, which couples super-resolution and detection through implicit feature sharing, saliency-guided focus, and gradient routing, addressing misalignment and feature redundancy issues in remote sensing detection.

Findings

Outperforms existing algorithms in small object detection accuracy.

Effectively suppresses background clutter and emphasizes weak object regions.

Maintains competitive computational efficiency.

Abstract

In remote sensing images, complex backgrounds, weak object signals, and small object scales make accurate detection particularly challenging, especially under low-quality imaging conditions. A common strategy is to integrate single-image super-resolution (SR) before detection; however, such serial pipelines often suffer from misaligned optimization objectives, feature redundancy, and a lack of effective interaction between SR and detection. To address these issues, we propose a Saliency-Driven multi-task Collaborative Network (SDCoNet) that couples SR and detection through implicit feature sharing while preserving task specificity. SDCoNet employs the swin transformer-based shared encoder, where hierarchical window-shifted self-attention supports cross-task feature collaboration and adaptively balances the trade-off between texture refinement and semantic representation. In addition, a multi-scale saliency prediction module produces importance scores to select key tokens, enabling focused attention on weak object regions, suppression of background clutter, and suppression of adverse features introduced by multi-task coupling. Furthermore, a gradient routing strategy is introduced to mitigate optimization conflicts. It first stabilizes detection semantics and subsequently routes SR gradients along a detection-oriented direction, enabling the framework to guide the SR branch to generate high-frequency details that are explicitly beneficial for detection. Experiments on public datasets, including NWPU VHR-10-Split, DOTAv1.5-Split, and HRSSD-Split, demonstrate that the proposed method, while maintaining competitive computational efficiency, significantly outperforms existing mainstream algorithms in small object detection on low-quality remote sensing images. Our code is available at https://github.com/qiruo-ya/SDCoNet.