D$^3$R-DETR: DETR with Dual-Domain Density Refinement for Tiny Object Detection in Aerial Images

TL;DR

D$^3$R-DETR introduces a dual-domain density refinement approach that fuses spatial and frequency information to improve tiny object detection accuracy in aerial images, addressing challenges of limited pixel info and density variation.

Contribution

It proposes a novel DETR-based detector with dual-domain density refinement, enhancing tiny object localization by leveraging spatial and frequency domain features.

Findings

Outperforms existing detectors on AI-TOD-v2 dataset

Refines low-level features for better density map prediction

Achieves faster convergence and higher accuracy

Abstract

Detecting tiny objects plays a vital role in remote sensing intelligent interpretation, as these objects often carry critical information for downstream applications. However, due to the extremely limited pixel information and significant variations in object density, mainstream Transformer-based detectors often suffer from slow convergence and inaccurate query-object matching. To address these challenges, we propose D$^3$R-DETR, a novel DETR-based detector with Dual-Domain Density Refinement. By fusing spatial and frequency domain information, our method refines low-level feature maps and utilizes their rich details to predict more accurate object density map, thereby guiding the model to precisely localize tiny objects. Extensive experiments on the AI-TOD-v2 dataset demonstrate that D$^3$R-DETR outperforms existing state-of-the-art detectors for tiny object detection.