GeoDexGrasp：面向数据高效与物理合理的几何感知灵巧抓取生成
GeoDexGrasp: Geometry-aware Generation for Data-efficient and Physics-plausible Dexterous Grasping
SIM(3) 等变网络加自监督解耦，把物体几何拆成形状、尺寸、位姿、交互方向的可迁移特征，这种以物体为中心的几何表征可直接迁移到视觉中的位姿估计与抓取检测，也可作为生成式抓取模型的条件编码器。
原文依据 We introduce a SIM(3)-equivariant network equipped with a self-supervised disentanglement strategy to extract interpretable and transferable geometric features
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论实现灵巧抓取仍是机器人领域的关键挑战。近期的生成式方法通过大规模数据驱动训练实现了多样化抓取，但往往忽视物体的几何先验，导致数据效率低、物理合理性差。我们提出 GeoDexGrasp，一个建立在以物体为中心的几何表征之上的几何感知灵巧抓取生成框架。我们引入一个 SIM(3) 等变网络，并配备自监督解耦策略，以提取可解释且可迁移的几何特征，包括形状、尺寸、位姿和交互方向。整个生成过程随后被分解为两个阶段：首先是以位姿和交互方向为条件的根节点旋转生成；其次是由形状和尺寸引导的手部抓取生成。借助几何表征，GeoDexGrasp 在五个数据集上达到 SOTA 的物理合理性（穿透深度降低 40%），并展现出更高的数据效率。此外，GeoDexGrasp 还十分轻量（参数量不足此前 SOTA 方法的 20%），并取得了可比的抓取成功率。
英文原文摘要 1176 字符
Achieving dexterous grasping remains a key challenge in robotics. Recent generative approaches enable diverse grasps through large-scale data-driven training, yet they often neglect geometric priors of objects, which leads to low data efficiency and poor physical plausibility. We propose GeoDexGrasp, a geometry-aware generation framework for dexterous grasping built upon object-centric geometric representations. We introduce a SIM(3)-equivariant network equipped with a self-supervised disentanglement strategy to extract interpretable and transferable geometric features, including shape, size, pose, and interaction direction. The overall generation process is then decomposed into two stages: first, root rotation generation conditioned on pose and interaction direction; second, hand grasp generation guided by shape and size. By leveraging geometric representations, GeoDexGrasp achieves SOTA physical plausibility (reducing 40% penetration depth) across five datasets, and exhibits improved data efficiency. Additionally, GeoDexGrasp is also lightweight (using less than 20% of the parameters of the previous SOTA method) and attains a comparable grasp success rate.