Knowledge combination to learn rotated detection without rotated annotation

T Zhu, B Ferenczi, P Purkait… - Proceedings of the …, 2023 - openaccess.thecvf.com
rotated detectors. In this work, we introduce Knowledge Combination to learn Rotated object
detection… axis-aligned annotation for the target dataset in order to predict rotated boxes. …

[PDF][PDF] Knowledge Combination to Learn Rotated Detection Without Rotated Annotation Supplementary Material

T Zhu, B Ferenczi, P Purkait, T Drummond… - openaccess.thecvf.com
… We include some banana images in Figure B1 to show why rotated detection is superior
to axis-aligned detection. It is also clear to see why rotated annotation is more expensive. …

Rotation-invariant features for multi-oriented text detection in natural images

C Yao, X Zhang, X Bai, W Liu, Y Ma, Z Tu - PloS one, 2013 - journals.plos.org
… of the authors’ knowledge, there is currently no related patent or … Hence, by combining the
strength of rotation-invariant … Besides, the ground truth rectangles annotated are not accurate …

Point2RBox: Combine Knowledge from Synthetic Visual Patterns for End-to-end Oriented Object Detection with Single Point Supervision

Y Yu, X Yang, Q Li, F Da, J Dai… - Proceedings of the …, 2024 - openaccess.thecvf.com
… object detection under the weak supervision of point annotations? Several point supervised
Knowledge combination to learn rotated detection without rotated annotation. In IEEE/CVF …

[PDF][PDF] H2rbox-v2: Boosting hbox-supervised oriented object detection via symmetric learning

Y Yu, X Yang, Q Li, Y Zhou, G Zhang… - arXiv preprint arXiv …, 2023 - researchgate.net
… Compared to H2RBox, our method is less susceptible to low annotation quality and …
Knowledge combination to learn rotated detection without rotated annotation. In IEEE/CVF …

H2RBox-v2: Incorporating symmetry for boosting horizontal box supervised oriented object detection

Y Yu, X Yang, Q Li, Y Zhou, F Da… - Advances in Neural …, 2024 - proceedings.neurips.cc
… very close performance to a rotation annotation trained counterpart – Rotated FCOS: 1) DOTA-…
Knowledge combination to learn rotated detection without rotated annotation. In IEEE/CVF …

PointOBB-v2: Towards Simpler, Faster, and Stronger Single Point Supervised Oriented Object Detection

B Ren, X Yang, Y Yu, J Luo, Z Deng - arXiv preprint arXiv:2410.08210, 2024 - arxiv.org
… introduce oriented detection supervised by rotated boxes. Then… We first utilize point annotations
for each object in the training … Knowledge combination to learn rotated detection without

A weak supervision learning paradigm for oriented ship detection in SAR image

T Yue, Y Zhang, J Wang, Y Xu… - … and Remote Sensing, 2024 - ieeexplore.ieee.org
… 1) To our knowledge, the paradigm we proposed is the first … detection, which requires only
the training of HBB annotation. … for the rotated targets detection of remote sensing, other HBB …

PointOBB: Learning Oriented Object Detection via Single Point Supervision

J Luo, X Yang, Y Yu, Q Li, J Yan… - Proceedings of the IEEE …, 2024 - openaccess.thecvf.com
… [35] combine HBox annotation and image rotation to align oriented objects horizontally or …
Knowledge combination to learn rotated detection without rotated annotation. In Proceedings …

PCBSSD: Self-supervised symmetry-aware detector for PCB displacement and orientation inspection

J Li, F Da, Y Yu - Measurement, 2025 - Elsevier
… their rotational symmetry … without any manual annotation, achieves performance comparable
to state-of-the-art supervised methods as to accuracy and speed. To our best knowledge