DextER: Language-driven Dexterous Grasp Generation with Embodied Reasoning
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_ | 1866910164856078336 |
|---|---|
| author | Lee, Junha Park, Eunha Cho, Minsu |
| author_facet | Lee, Junha Park, Eunha Cho, Minsu |
| contents | Language-driven dexterous grasp generation requires the models to understand task semantics, 3D geometry, and complex hand-object interactions. While vision-language models have been applied to this problem, existing approaches directly map observations to grasp parameters without intermediate reasoning about physical interactions. We present DextER, Dexterous Grasp Generation with Embodied Reasoning, which introduces contact-based embodied reasoning for multi-finger manipulation. Our key insight is that predicting which hand links contact where on the object surface provides an embodiment-aware intermediate representation, bridging task semantics with physical constraints. DextER autoregressively generates embodied contact tokens specifying which finger links contact where on the object surface, followed by grasp tokens encoding the hand configuration. On DexGYS, DextER achieves 67.14% success rate, outperforming state-of-the-art by 3.83 p.p. with 96.4% improvement in intention alignment. We also demonstrate steerable generation through partial contact specification, providing fine-grained control over grasp synthesis. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_16046 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DextER: Language-driven Dexterous Grasp Generation with Embodied Reasoning Lee, Junha Park, Eunha Cho, Minsu Robotics Computer Vision and Pattern Recognition Language-driven dexterous grasp generation requires the models to understand task semantics, 3D geometry, and complex hand-object interactions. While vision-language models have been applied to this problem, existing approaches directly map observations to grasp parameters without intermediate reasoning about physical interactions. We present DextER, Dexterous Grasp Generation with Embodied Reasoning, which introduces contact-based embodied reasoning for multi-finger manipulation. Our key insight is that predicting which hand links contact where on the object surface provides an embodiment-aware intermediate representation, bridging task semantics with physical constraints. DextER autoregressively generates embodied contact tokens specifying which finger links contact where on the object surface, followed by grasp tokens encoding the hand configuration. On DexGYS, DextER achieves 67.14% success rate, outperforming state-of-the-art by 3.83 p.p. with 96.4% improvement in intention alignment. We also demonstrate steerable generation through partial contact specification, providing fine-grained control over grasp synthesis. |
| title | DextER: Language-driven Dexterous Grasp Generation with Embodied Reasoning |
| topic | Robotics Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.16046 |