Answers are grounded in this paper and cite their sources.
Rotated object detection plays a vital role in remote sensing interpretation, with broad applications in urban planning, port monitoring, and disaster response. However, the significant scale variations, complex orientations, and cluttered backgrounds in remote sensing images pose considerable challenges to accurate detection. To address these issues, this paper proposes an efficient rotated object detection framework that integrates state space models with vision transformers to achieve an optimal balance between accuracy and computational efficiency.The proposed framework employs a MobileMamba backbone enhanced with a Multi-Receptive Field Feature Interaction (MRFFI) module for effective local-global feature representation. An EfficientViT-FPN neck enables efficient multi-scale feature fusion, while a refined Rotated RetinaNet head incorporates five-parameter rotated box regression with angle-aware constraints to improve orientation estimation. Comprehensive experiments on the DOTA-v1.0 and SRSDD-v1.0 datasets demonstrate that our approach achieves superior detection accuracy with significantly reduced computational overhead, making it particularly suitable for practical remote sensing applications
No takes yet. Share an insight, caveat, or question.
Lin et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: