Bi-Adapt: Few-shot Bimanual Adaptation for Novel Categories of 3D Objects via Semantic Correspondence
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917262685896704 |
|---|---|
| author | Zhou, Jinxian Wu, Ruihai Liu, Yiwei Hou, Yiwen Zhou, Xunzhe Yu, Checheng Zhong, Licheng Shao, Lin |
| author_facet | Zhou, Jinxian Wu, Ruihai Liu, Yiwei Hou, Yiwen Zhou, Xunzhe Yu, Checheng Zhong, Licheng Shao, Lin |
| contents | Bimanual manipulation is imperative yet challenging for robots to execute complex tasks, requiring coordinated collaboration between two arms. However, existing methods for bimanual manipulation often rely on costly data collection and training, struggling to generalize to unseen objects in novel categories efficiently. In this paper, we present Bi-Adapt, a novel framework designed for efficient generalization for bimanual manipulation via semantic correspondence. Bi-Adapt achieves cross-category affordance mapping by leveraging the strong capability of vision foundation models. Fine-tuning with restricted data on novel categories, Bi-Adapt exhibits notable generalization to out-of-category objects in a zero-shot manner. Extensive experiments conducted in both simulation and real-world environments validate the effectiveness of our approach and demonstrate its high efficiency, achieving a high success rate on different benchmark tasks across novel categories with limited data. Project website: https://biadapt-project.github.io/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08425 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Bi-Adapt: Few-shot Bimanual Adaptation for Novel Categories of 3D Objects via Semantic Correspondence Zhou, Jinxian Wu, Ruihai Liu, Yiwei Hou, Yiwen Zhou, Xunzhe Yu, Checheng Zhong, Licheng Shao, Lin Robotics Bimanual manipulation is imperative yet challenging for robots to execute complex tasks, requiring coordinated collaboration between two arms. However, existing methods for bimanual manipulation often rely on costly data collection and training, struggling to generalize to unseen objects in novel categories efficiently. In this paper, we present Bi-Adapt, a novel framework designed for efficient generalization for bimanual manipulation via semantic correspondence. Bi-Adapt achieves cross-category affordance mapping by leveraging the strong capability of vision foundation models. Fine-tuning with restricted data on novel categories, Bi-Adapt exhibits notable generalization to out-of-category objects in a zero-shot manner. Extensive experiments conducted in both simulation and real-world environments validate the effectiveness of our approach and demonstrate its high efficiency, achieving a high success rate on different benchmark tasks across novel categories with limited data. Project website: https://biadapt-project.github.io/ |
| title | Bi-Adapt: Few-shot Bimanual Adaptation for Novel Categories of 3D Objects via Semantic Correspondence |
| topic | Robotics |
| url | https://arxiv.org/abs/2602.08425 |