Fourier Angle Alignment for Oriented Object Detection in Remote Sensing

TL;DR

This paper introduces Fourier Angle Alignment, a novel method leveraging Fourier rotation equivariance to improve oriented object detection in remote sensing images, achieving state-of-the-art results.

Contribution

It proposes Fourier Angle Alignment and two plug-and-play modules, FAAFusion and FAA Head, to address directional incoherence and task conflict in remote sensing object detection.

Findings

Achieves 78.72% mAP on DOTA-v1.0

Achieves 72.28% mAP on DOTA-v1.5

Significantly improves detection performance

Abstract

In remote sensing rotated object detection, mainstream methods suffer from two bottlenecks, directional incoherence at detector neck and task conflict at detecting head. Ulitising fourier rotation equivariance, we introduce Fourier Angle Alignment, which analyses angle information through frequency spectrum and aligns the main direction to a certain orientation. Then we propose two plug and play modules : FAAFusion and FAA Head. FAAFusion works at the detector neck, aligning the main direction of higher-level features to the lower-level features and then fusing them. FAA Head serves as a new detection head, which pre-aligns RoI features to a canonical angle and adds them to the original features before classification and regression. Experiments on DOTA-v1.0, DOTA-v1.5 and HRSC2016 show that our method can greatly improve previous work. Particularly, our method achieves new state-of-the-art results of 78.72% mAP on DOTA-v1.0 and 72.28% mAP on DOTA-v1.5 datasets with single scale training and testing, validating the efficacy of our approach in remote sensing object detection. The code is made publicly available at https://github.com/gcy0423/Fourier-Angle-Alignment .