S2C-Net: SAM2-Based Dual-Domain Feature Reconstruction and Semantic Decoupling for Tiny Remote Sensing Object Counting (opens in new tab)

Remote sensing object counting (RSOC) is a fundamental task with significant potential applications in urban traffic monitoring, road planning, and public security. However, accurate counting remains challenging since tiny objects are easily confused with complex backgrounds in remote sensing (RS) scenarios. Recent vision foundation models (VFMs) have brought new opportunities to RSOC owing to their powerful ability to learn general representations. Nevertheless, VFMs cannot be directly appli...