PointOBB-v3： Expanding Performance Boundaries of Single Point-Supervised Oriented Object Detection

Peiyuan Zhang  Junwei Luo  Xue Yang  Yi Yu  Qingyun Li  Yue Zhou  Xiaosong Jia  Xudong Lu  Jingdong Chen  Xiang Li  Junchi Yan  Yansheng Li 

If you find our work helpful, please consider giving us a ⭐!

The paper is available at PointOBB-v3. You are also welcome to check out the conference version PointOBB.

📌 Note: This branch contains the code for the two-stage version. For the end-to-end version, please refer to end-to-end branch.

Train/Test

Please see PointOBB/README.md.

Weight

DIOR-R

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	41.82	le90	pointobbv3-dior	Oriented RCNN	model

DOTA-v1.0

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	50.44	le90	pointobbv3-dota	Oriented RCNN	model

DOTA-v1.5

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	38.08	le90	pointobbv3-dota15	Oriented RCNN	model

DOTA-v2.0

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	24.86	le90	pointobbv3-dota20	Oriented RCNN	model

FAIR1M

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	20.19	le90	pointobbv3-fair	Oriented RCNN	model

STAR

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	16.73	le90	pointobbv3-star	Oriented RCNN	model

RSAR

Backbone	mAP	Angle	Config	Detector	Download
ResNet50 (1024,1024,200)	22.84	le90	pointobbv3-rsar	Oriented RCNN	model

Citation

If you find this work helpful, please consider to cite:

@article{zhang2025pointobb,
   title     = {PointOBB-v3: Expanding Performance Boundaries of Single Point-Supervised Oriented Object Detection},
   author     = {Zhang, Peiyuan and Luo, Junwei and Yang, Xue and Yu, Yi and Li, Qingyun and Zhou, Yue and Jia, Xiaosong and Lu, Xudong and Chen, Jingdong and Li, Xiang and others},
   journal    = {arXiv preprint arXiv:2501.13898},
   year       = {2025}
}

@InProceedings{luo2024pointobb,
   title     = {PointOBB: Learning Oriented Object Detection via Single Point Supervision},
   author    = {Luo, Junwei and Yang, Xue and Yu, Yi and Li, Qingyun and Yan, Junchi and Li, Yansheng},
   booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
   pages     = {16730-16740},
   year      = {2024}
}

---

Special thanks to the codebase contributors of MMRotate and P2BNet!

@inproceedings{zhou2022mmrotate,
  title   = {MMRotate: A Rotated Object Detection Benchmark using PyTorch},
  author  = {Zhou, Yue and Yang, Xue and Zhang, Gefan and Wang, Jiabao and Liu, Yanyi and
             Hou, Liping and Jiang, Xue and Liu, Xingzhao and Yan, Junchi and Lyu, Chengqi and
             Zhang, Wenwei and Chen, Kai},
  booktitle={Proceedings of the 30th ACM International Conference on Multimedia},
  year={2022}
}

@inproceedings{P2BNet,
  title     = {Point-to-Box Network for Accurate Object Detection via Single Point Supervision},
  author    = {Pengfei Chen, Xuehui Yu, Xumeng Han, Najmul Hassan, Kai Wang, Jiachen Li, Jian Zhao, Humphrey Shi, Zhenjun Han, and Qixiang Ye},
  booktitle = {ECCV},
  year      = {2022}
}