GRAFT: 基于部件对应的图结构功能迁移
GRAFT: Graph-Based Affordance Transfer via Part Correspondence
AI总结：
提出GRAFT框架，利用部件级图表示和几何感知对应，通过单次演示实现零样本操作迁移，解决泛化到未见物体的挑战。
摘要：
将机器人操作泛化到未见物体仍然具有挑战性，因为基于学习的方法需要大量演示并在少样本设置中失败。先前的工作通过语义检索迁移功能，但仅凭语义忽略了对于操作至关重要的几何相似性。我们提出GRAFT，一种几何感知的对应框架，每个物体仅需一次演示即可实现零样本操作迁移。物体被表示为基于部件的图，其中部件级描述符支持全局实例检索和部件对应，顶点级描述符支持细粒度接触点匹配。对于未见物体，我们的方法首先从演示缓冲区中检索功能上和几何上最相似的实例，并对其功能部件进行对齐，最后通过逐点对应传播接触点。
英文摘要：
Generalizing robotic manipulation to unseen objects remains challenging, as learning-based approaches require many demonstrations and fail in few-shot settings. Prior work transfers affordances through semantic retrieval, but semantics alone neglect geometric similarity, which is critical for manipulation. We propose GRAFT, a geometry-aware correspondence framework for zero-shot manipulation transfer using only one demonstration per object. Objects are represented as part-based graphs, where part-level descriptors support global instance retrieval and part correspondence, and vertex-level descriptors enable fine-grained contact point matching. For an unseen object, our method first retrieves the most functionally and geometrically similar instance from the demonstration buffer with aligned functional parts, and finally propagates the contact points through point-wise correspondence.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
