面向未知关节物体数字孪生构建的神经隐式表示
Neural Implicit Representation for Building Digital Twins of Unknown Articulated Objects
AI总结：
本文提出一种从两次RGBD扫描构建未知关节物体数字孪生的方法，通过两阶段重建物体形状并恢复部件分割与关节运动，显式建模点级对应关系，实现更准确稳定的结果且无需形状先验。
摘要：
我们解决从物体在不同关节状态下的两次RGBD扫描构建未知关节物体数字孪生的问题。我们将该问题分解为两个阶段，每个阶段处理不同的方面。我们的方法首先重建每个状态下的物体级形状，然后恢复底层关节模型，包括部件分割和关联两个状态的关节运动。通过显式建模点级对应关系，并利用来自图像、三维重建和运动学的线索，与先前工作相比，我们的方法产生了更准确和更稳定的结果。它还能处理多个可动部件，并且不依赖任何物体形状或结构先验。项目页面：https://github.com/NVlabs/DigitalTwinArt
英文摘要：
We address the problem of building digital twins of unknown articulated objects from two RGBD scans of the object at different articulation states. We decompose the problem into two stages, each addressing distinct aspects. Our method first reconstructs object-level shape at each state, then recovers the underlying articulation model including part segmentation and joint articulations that associate the two states. By explicitly modeling point-level correspondences and exploiting cues from images, 3D reconstructions, and kinematics, our method yields more accurate and stable results compared to prior work. It also handles more than one movable part and does not rely on any object shape or structure priors. Project page: https://github.com/NVlabs/DigitalTwinArt
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。
