🤖 AI Summary
To address orientation-sensitive ship detection in SAR imagery, this paper proposes the first end-to-end framework for oriented object detection based on Sparse Learnable Proposals (SLP). Departing from conventional dense anchor-based paradigms, our method employs only 300 learnable, oriented proposals and jointly optimizes their initialization, oriented bounding box (OBB) regression, and classification. We adapt the Sparse R-CNN paradigm to rotation-aware detection for the first time, explicitly modeling rotational parameters—enabling direct OBB output without non-maximum suppression (NMS) or post-processing. After fine-tuning on the RSDD-SAR dataset, our approach surpasses existing state-of-the-art methods in both near-shore and open-sea scenarios, achieving significant AP gains. Moreover, it yields a more lightweight model and exhibits enhanced training stability. The source code is publicly available.
📝 Abstract
We present Sparse R-CNN OBB, a novel framework for the detection of oriented objects in SAR images leveraging sparse learnable proposals. The Sparse R-CNN OBB has streamlined architecture and ease of training as it utilizes a sparse set of 300 proposals instead of training a proposals generator on hundreds of thousands of anchors. To the best of our knowledge, Sparse R-CNN OBB is the first to adopt the concept of sparse learnable proposals for the detection of oriented objects, as well as for the detection of ships in Synthetic Aperture Radar (SAR) images. The detection head of the baseline model, Sparse R-CNN, is re-designed to enable the model to capture object orientation. We also fine-tune the model on RSDD-SAR dataset and provide a performance comparison to state-of-the-art models. Experimental results shows that Sparse R-CNN OBB achieves outstanding performance, surpassing other models on both inshore and offshore scenarios. The code is available at: www.github.com/ka-mirul/Sparse-R-CNN-OBB.