Introduces SCOUT, a routing framework that intelligently selects which Image-to-3D reconstruction model to use based on input difficulty and cost constraints.
March 31, 2026
Original Paper
Which Reconstruction Model Should a Robot Use? Routing Image-to-3D Models for Cost-Aware Robotic Manipulation
arXiv · 2603.27797
The Takeaway
Critical for cost-constrained robotics, this method allows a system to dynamically choose between 'cheap/fast' and 'expensive/high-quality' 3D models. It treats model selection as a learned routing problem, optimizing the tradeoff between reconstruction fidelity and latency/compute cost.
From the abstract
Robotic manipulation tasks require 3D mesh reconstructions of varying quality: dexterous manipulation demands fine-grained surface detail, while collision-free planning tolerates coarser representations. Multiple reconstruction methods offer different cost-quality tradeoffs, from Image-to-3D models - whose output quality depends heavily on the input viewpoint - to view-invariant methods such as structured light scanning. Querying all models is computationally prohibitive, motivating per-input mo