# Robo3R: Enhancing Robotic Manipulation with Accurate Feed-Forward 3D Reconstruction

arXiv: https://arxiv.org/abs/2602.10101 · HTML: https://mlsanity.com/paper/2602.10101/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Feb 2026.

## Built on by (most cited first, 9 of 9)

- [Evo-Depth: A Lightweight Depth-Enhanced Vision-Language-Action Model](https://mlsanity.com/paper/2605.14950/) (2026)
- [Lift3D-VLA: Lifting VLA Models to 3D Geometry and Dynamics-Aware Manipulation](https://mlsanity.com/paper/2607.06564/) (2026)
- [Spark3R: Asymmetric Token Reduction Makes Fast Feed-Forward 3D Reconstruction](https://mlsanity.com/paper/2605.06270/) (2026)
- [Adaptive Vision-Language Grasping via Composable Foundation Priors and Generalizable Grasp Synthesis](https://mlsanity.com/paper/2609.04096/) (2026)
- [Not All Tasks Quantize Equally: Fisher-Guided Quantization for Visual Geometry Transformer](https://mlsanity.com/paper/2605.15828/) (2026)
- [SpatialBench: Is Your Spatial Foundation Model an All-Round Player?](https://mlsanity.com/paper/2605.27367/) (2026)
- [DarkVGGT: Seeing Through Darkness Using Thermal Geometry](https://mlsanity.com/paper/2606.11326/) (2026)
- [RoboInter1.5: A Holistic Intermediate Representation Suite for Embodied World Modeling and Robotic Manipulation](https://mlsanity.com/paper/2607.18709/) (2026)
- [3DROID: A Renderable 3D Gaussian Dataset with Measured Per-Scene Reliability](https://mlsanity.com/paper/2610.01744/) (2026)

## Abstract (arXiv)

3D spatial perception is fundamental to generalizable robotic manipulation, yet obtaining reliable, high-quality 3D geometry remains challenging. Depth sensors suffer from noise and material sensitivity, while existing reconstruction models lack the precision and metric consistency required for physical interaction. We introduce Robo3R, a feed-forward, manipulation-ready 3D reconstruction model that predicts accurate, metric-scale scene geometry directly from RGB images and robot states in real time. Robo3R jointly infers scale-invariant local geometry and relative camera poses, which are unified into the scene representation in the canonical robot frame via a learned global similarity transformation. To meet the precision demands of manipulation, Robo3R employs a masked point head for sharp, fine-grained point clouds, and a keypoint-based Perspective-n-Point (PnP) formulation to refine camera extrinsics and global alignment. Trained on Robo3R-4M, a curated large-scale synthetic dataset with four million high-fidelity annotated frames, Robo3R consistently outperforms state-of-the-art reconstruction methods and depth sensors. Across downstream tasks including imitation learning, sim-to-real transfer, grasp synthesis, and collision-free motion planning, we observe consistent gains in performance, suggesting the promise of this alternative 3D sensing module for robotic manipulation.