Scale-Aware Swin Transformer with Adaptive Pyramid Fusion for Tiny Object Detection in Aerial Images

Author(s)	Priyanka Sahani, Dr. Ajay Singh
Country	India
Abstract	Detecting aerial objects in remote sensing images is difficult due to significant scale differences, dense arrangements, random orientations, and messy backgrounds. Traditional convolutional detectors frequently overlook small targets and have difficulty in capturing long-range contextual relationships. This paper introduces a Scale-Aware Swin Transformer with Adaptive Pyramid Fusion (SST-APF) aimed at multi-scale detection of aerial objects. The architecture integrates a hierarchical transformer backbone, adaptive feature pyramid fusion, and a scale-sensitive attention refinement module to enhance the detection of small, medium, and large objects. Experimental assessment on DOTA, VisDrone, and xView shows reliable improvements over CNN and transformer benchmarks in mean average precision and small-object recall. The suggested architecture is appropriate for monitoring, traffic evaluation, disaster assessment, and smart city oversight.
Keywords	Aerial object detection, vision transformer, remote sensing, UAV imagery, multi-scale learning.
Field	Computer Applications
Published In	Volume 7, Issue 3, May-June 2026
Published On	2026-05-05

About AIJFR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9898-79-39-59	editor@aijfr.com

Advanced International Journal for Research