arXiv AI
YUBI-Real2Sim: 3DGS Scene Verification for Reliable Policy Evaluation
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv AI
Sep 23
ReVeal: A Reconstruction-Aware Real-to-Sim Framework for VLA Policy Evaluation
arXiv:2609.23910v1 Announce Type: cross Abstract: Simulation-based evaluation provides a scalable and repeatable alternative to real-world evaluation of vision-language-action (VLA) policies. However...
arXiv AI
Jun 4
VISTA: Vision-Grounded and Physics-Validated Adaptation of UMI data for VLA Training
arXiv:2606. 04708v1 Announce Type: cross Abstract: Universal Manipulation Interface (UMI) enables scalable real-world robot data collection without hardware-specific teleoperation, yet leveraging UMI data to train large-scale Vision-Language-Action (VLA) models remains fundamentally challenging.
arXiv Computer Vision
Aug 27
PIVOT: A Multi-Trajectory Dataset and Testbed for Pose, Intrinsics, and Novel Viewpoint Evaluation in Real-World 3D Reconstruction
PIVOT is a new multi‑trajectory dataset and evaluation framework that captures real‑world scenes with diverse camera paths, preserving both sensor‑derived measured poses and COLMAP‑optimized poses along with calibrated and optimized intrinsics. It defines three benchmark families—seen vs. unseen trajectory generalization, measured vs. optimized pose sensitivity, and calibrated vs. optimized intrinsics sensitivity—and introduces a directed pose‑space Chamfer distance to assess pose coverage. The first version of PIVOT includes five scenes recorded with a DJI Mini 4 Pro and offers an open processing and Nerfstudio‑based evaluation toolchain, revealing a consistent quality gap between held‑out and unseen trajectories and significant sensitivity to pose source and camera intrinsics.
arXiv AI
Sep 1
RoboPhys-3D: A Comprehensive Embodied World Model Evaluation via 3D Reconstruction
RoboPhys-3D is a 3D‑grounded embodied world model benchmark built on RoboTwin 2.0, featuring 50 manipulation tasks, 5,000 episodes, and 25,000 multi‑view ground‑truth videos. It evaluates video world models by processing both generated and ground‑truth videos through the same 3D reconstruction pipeline, allowing the separation of reconstruction‑induced from generation‑induced errors. The benchmark defines 50 metrics across four sub‑dimensions—pixel fidelity, 3D geometry consistency, state understanding, and task completeness—and introduces the Average Full Score and RoboPhyscore for holistic assessment, with RoboPhyscore showing strong correlation with human judgments.
Hugging Face Trending Papers
Jun 3
VISTA: Vision-Grounded and Physics-Validated Adaptation of UMI data for VLA Training
Universal Manipulation Interface (UMI) enables scalable real-world robot data collection without hardware-specific teleoperation, yet leveraging UMI data to train large-scale Vision-Language-Action (VLA) models remains fundamentally challenging. We identify two critical mismatches: wrist-mounted fisheye views, with severe radial distortion and local gripper-centric perspectives, are out-of-distribution for pretrained VLMs; and human-collected trajectories frequently violate kinematic limits, incur collisions, or exceed controller bandwidth, teaching VLA policies physically infeasible actions.
arXiv Computer Vision
Oct 2
3DROID: A Renderable 3D Gaussian Dataset with Measured Per-Scene Reliability
The paper introduces 3DROID, a dataset of renderable 3D Gaussian scenes that are anchored to a robot’s metric workspace and include per-scene reliability metrics. It examines how the reliability of camera extrinsics and pose conditioning affect the fidelity of 3D Gaussian representations, proposing a calibration-aware pipeline that improves novel-view rendering when extrinsics are trustworthy. The resulting dataset, available on Hugging Face, provides robot manipulation researchers with metric-scale, pose-anchored 3D data and reliability annotations.