Is Single-View Mesh Reconstruction Ready for Robotics?
Nolte, Geiger, Schölkopf, Posner · cs.RO,cs.CV · 2026-09-05 · 原文
This paper evaluates single-view mesh reconstruction models for their potential in enabling instant digital twin creation for real-time planning and dynamics prediction using physics simulators for robotic manipulation. Recent single-view 3D reconstruction advances offer a promising avenue toward an automated real-to-sim pipeline: directly mapping a single observation of a scene into a simulation instance by reconstructing scene objects as individual, complete, and physically plausible 3D meshes. However, their suitability for physics simulations and robotics applications under immediacy, physical fidelity, and simulation readiness remains underexplored. We establish robotics-specific benchmarking criteria for 3D reconstruction, including handling typical inputs, collision-free and stable geometry, occlusion robustness, and meeting computational constraints. Our empirical evaluation using realistic robotics datasets shows that despite success on computer vision benchmarks, existing approaches fail to meet robotics-specific requirements. We quantitively examine limitations of single-view reconstruction for practical robotics implementation, in contrast to prior work that focuses on
讲义
讲义·推断 依据「原文」自动生成的结构化摘要（推断），非原文表述；以原文为准。
1. 人话版
This paper evaluates single-view mesh reconstruction models for their potential in enabling instant digital twin creation for real-time planning and dynamics prediction using physics simulators for robotic manipulation.
Recent single-view 3D reconstruction advances offer a promising avenue toward an automated real-to-sim pipeline: directly mapping a single observation of a scene into a simulation instance by reconstructing scene objects as individual, complete, and physically plausible 3D meshes.
2. 领域脉络
本文类目：cs.RO、cs.CV，属于其所在研究脉络的最新进展。
3. 机制拆解
摘要未展开方法细节——精读时重点看方法/模型部分。
4. 证据与数字
We establish robotics-specific benchmarking criteria for 3D reconstruction, including handling typical inputs, collision-free and stable geometry, occlusion robustness, and meeting computational constraints.
5. 反例与边界
However, their suitability for physics simulations and robotics applications under immediacy, physical fidelity, and simulation readiness remains underexplored.
Our empirical evaluation using realistic robotics datasets shows that despite success on computer vision benchmarks, existing approaches fail to meet robotics-specific requirements.
We quantitively examine limitations of single-view reconstruction for practical robotics implementation, in contrast to prior work that focuses on
6. 跨领域连接与意外收获
横跨 2 个类目（cs.RO、cs.CV），关注其在你兴趣板块间的迁移面。
7. 可复用方法
把本文机制与你手头项目对照，找一个两周内能验证的最小实验。
8. 术语表
精读时把不熟的术语记入此处，作为下次回忆的锚点。