Details

Autor(en) / Beteiligte

• Huang, Qiangqiang
• Yao, Ruilin
• Lu, Xiaoqiang
• Zhu, Jishuai
• Xiong, Shengwu
• Chen, Yaxiong

Titel

Oriented Object Detector With Gaussian Distribution Cost Label Assignment and Task-Decoupled Head

Ist Teil von

• IEEE transactions on geoscience and remote sensing, 2024, Vol.62, p.1-16

Ort / Verlag

New York: IEEE

Erscheinungsjahr

2024

Quelle

IEEE Electronic Library (IEL)

Beschreibungen/Notizen

• Recently, oriented object detection in remote sensing images has garnered significant attention due to its broad range of applications. Early-oriented object detection adhered to the established general object detection frameworks, utilizing the label assignment strategy based on the horizontal bounding box (HBB) annotations or rotation-agnostic cost function. Such a strategy may not reflect the large aspect ratio and rotation of arbitrary-oriented objects in remote sensing images and require high parameter-tuning efforts in the training process, which will eventually harm the detector performance. Furthermore, the localization quality of oriented objects depends on precise rotation angle prediction, exacerbating the inconsistency between classification and regression tasks in oriented object detection. To address these issues, we propose the Gaussian distribution cost optimal transport assignment (GCOTA) and decoupled layer attention angle head (DLAAH). Specifically, GCOTA utilizes a Gaussian distribution-based cost function for the optimal transport (OT) label assignment in the training process, alleviating the impact of rotation angle and large aspect ratio in remote sensing images. DLAAH predicts rotation angle independently and incorporates layer attention to obtain the task-specific features based on the shared FPN features, enhancing the angle prediction and improving consistency across different tasks. Based on these proposed components, we present an anchor-free oriented detector, namely, Gaussian distribution and task-decoupled head oriented detector (GTDet) and a multiclass ship detection dataset in real scenarios (CGWX), which provides a benchmark for fine-grained object recognition in remote sensing images. Comprehensive experiments are conducted on CGWX and several public challenging datasets, including DOTAv1.0, and HRSC2016, to demonstrate that our method achieves superior performance on oriented object detection tasks. The code is available at https://github.com/WUTCM-Lab/GTDet .