Generalizable and Actionable Parts Pose Estimation with Symmetry Annotation-Free Learning Strategy

Wenxiao Chen, Xueyu Yuan, Liu Liu, Di Wu, Dan Guo
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:17960-17978, 2026.

Abstract

Urgently needed generalizable robot object interaction and manipulation requires high-quality Cross-Category object perception. As a pioneer of this area, Generalizable and Actionable Parts (GAParts) understanding has attracted increasing attention from relevant researchers. However, most recent works either have insufficient design regarding the symmetry issue or require rich symmetry annotation, which severely impedes precise GAPart pose estimation in data-lacking scenarios. In this paper, we propose SAFAG, a novel Symmetry Annotation-Free framework for Generalizable and Actionable Parts Pose Estimation. Specifically, we suggest a stepwise refinement two-stage framework for candidate-to-final quaternion regression, and tackle the symmetry prediction as a probability distribution problem with self-supervised learning strategy. The experimental results demonstrate the superior performance and robustness of our SAFAG. We believe that our work has the enormous potential to be applied in many areas of embodied AI system.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-chen26fw, title = {Generalizable and Actionable Parts Pose Estimation with Symmetry Annotation-Free Learning Strategy}, author = {Chen, Wenxiao and Yuan, Xueyu and Liu, Liu and Wu, Di and Guo, Dan}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {17960--17978}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/chen26fw/chen26fw.pdf}, url = {https://proceedings.mlr.press/v306/chen26fw.html}, abstract = {Urgently needed generalizable robot object interaction and manipulation requires high-quality Cross-Category object perception. As a pioneer of this area, Generalizable and Actionable Parts (GAParts) understanding has attracted increasing attention from relevant researchers. However, most recent works either have insufficient design regarding the symmetry issue or require rich symmetry annotation, which severely impedes precise GAPart pose estimation in data-lacking scenarios. In this paper, we propose SAFAG, a novel Symmetry Annotation-Free framework for Generalizable and Actionable Parts Pose Estimation. Specifically, we suggest a stepwise refinement two-stage framework for candidate-to-final quaternion regression, and tackle the symmetry prediction as a probability distribution problem with self-supervised learning strategy. The experimental results demonstrate the superior performance and robustness of our SAFAG. We believe that our work has the enormous potential to be applied in many areas of embodied AI system.} }
Endnote
%0 Conference Paper %T Generalizable and Actionable Parts Pose Estimation with Symmetry Annotation-Free Learning Strategy %A Wenxiao Chen %A Xueyu Yuan %A Liu Liu %A Di Wu %A Dan Guo %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-chen26fw %I PMLR %P 17960--17978 %U https://proceedings.mlr.press/v306/chen26fw.html %V 306 %X Urgently needed generalizable robot object interaction and manipulation requires high-quality Cross-Category object perception. As a pioneer of this area, Generalizable and Actionable Parts (GAParts) understanding has attracted increasing attention from relevant researchers. However, most recent works either have insufficient design regarding the symmetry issue or require rich symmetry annotation, which severely impedes precise GAPart pose estimation in data-lacking scenarios. In this paper, we propose SAFAG, a novel Symmetry Annotation-Free framework for Generalizable and Actionable Parts Pose Estimation. Specifically, we suggest a stepwise refinement two-stage framework for candidate-to-final quaternion regression, and tackle the symmetry prediction as a probability distribution problem with self-supervised learning strategy. The experimental results demonstrate the superior performance and robustness of our SAFAG. We believe that our work has the enormous potential to be applied in many areas of embodied AI system.
APA
Chen, W., Yuan, X., Liu, L., Wu, D. & Guo, D.. (2026). Generalizable and Actionable Parts Pose Estimation with Symmetry Annotation-Free Learning Strategy. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:17960-17978 Available from https://proceedings.mlr.press/v306/chen26fw.html.

Related Material