DASSF: Dynamic-Attention Scale-Sequence Fusion for Aerial Object Detection

TL;DR

This paper introduces DASSF, a novel method combining dynamic scale-sequence fusion, a small object detection head, and dynamic attention to enhance aerial small object detection accuracy, especially for overlapping and fuzzy targets.

Contribution

It proposes a universal framework that improves small object detection in aerial images by integrating dynamic attention, scale fusion, and a dedicated small object detection head into YOLO-based models.

Findings

DASSF improves mAP by 9.2% on VisDrone-2019 dataset.

DASSF enhances detection of small and overlapping objects.

The method outperforms current mainstream detection approaches.

Abstract

The detection of small objects in aerial images is a fundamental task in the field of computer vision. Moving objects in aerial photography have problems such as different shapes and sizes, dense overlap, occlusion by the background, and object blur, however, the original YOLO algorithm has low overall detection accuracy due to its weak ability to perceive targets of different scales. In order to improve the detection accuracy of densely overlapping small targets and fuzzy targets, this paper proposes a dynamic-attention scale-sequence fusion algorithm (DASSF) for small target detection in aerial images. First, we propose a dynamic scale sequence feature fusion (DSSFF) module that improves the up-sampling mechanism and reduces computational load. Secondly, a x-small object detection head is specially added to enhance the detection capability of small targets. Finally, in order to improve the expressive ability of targets of different types and sizes, we use the dynamic head (DyHead). The model we proposed solves the problem of small target detection in aerial images and can be applied to multiple different versions of the YOLO algorithm, which is universal. Experimental results show that when the DASSF method is applied to YOLOv8, compared to YOLOv8n, on the VisDrone-2019 and DIOR datasets, the model shows an increase of 9.2% and 2.4% in the mean average precision (mAP), respectively, and outperforms the current mainstream methods.