引用与许可
论文
来源与许可
项目许可证见 LICENSE。公开代码和素材来源见来源清单(Markdown);WAPR 接口来源于 WAPR。第三方组件遵循各自的许可证,并保留原作者声明。
默认 RGB 检测器使用 Ultralytics。部署前请按实际用途核对 Ultralytics 许可,并确认所用检测器和位姿权重的授权范围。WAPR 的四份第一方权重遵循单独的权重许可。
网站三维示意使用 Three.js,按 MIT 许可证分发。
BibTeX 引用
@InProceedings{HccePose_BF,
author = {Wang, Yulin and Hu, Mengting and Li, Hongli and Luo, Chen},
title = {HccePose(BF): Predicting Front \& Back Surfaces to Construct Ultra-Dense 2D-3D Correspondences for Pose Estimation},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
month = {October},
year = {2025},
pages = {7166-7175}
}
致谢
- BOP 与 BOP toolkit 提供数据格式、基准与评测工具。
- BlenderProc 用于训练场景渲染,KASAL 用于旋转对称信息。
- Ultralytics YOLO 提供检测框。
- 可选 RGB-D 接入使用公开 WAPR 和 MegaPose。
Citation and licenses
Paper
Sources and licenses
See the project license and the source list (Markdown) for published code and media sources. The WAPR interface comes from WAPR. Third-party components retain their own licenses and author notices.
The default RGB detector uses Ultralytics. Check the Ultralytics licensing options for your intended use and the terms for the checkpoints you obtain separately. The four first-party WAPR checkpoints have their own weight license.
The 3D illustration uses Three.js under MIT.
BibTeX citation
@InProceedings{HccePose_BF,
author = {Wang, Yulin and Hu, Mengting and Li, Hongli and Luo, Chen},
title = {HccePose(BF): Predicting Front \& Back Surfaces to Construct Ultra-Dense 2D-3D Correspondences for Pose Estimation},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
month = {October},
year = {2025},
pages = {7166-7175}
}
Acknowledgments
- BOP and BOP toolkit provide formats, benchmarks and evaluation tools.
- BlenderProc supports training-scene rendering; KASAL describes rotational symmetries.
- Ultralytics YOLO provides detections.
- Optional RGB-D integration uses public WAPR and MegaPose.