analiza aiqbench
This paper uses learned models, including distilled 3D vision foundation model features and a custom training objective with hard mining, to synthesize realistic depth-sensor noise for sim-to-real robotics. The reported 92.1% real-world success rate comes from the authors' own robotic experiments and needs independent replication, but the core idea of separating sensing dropouts from measurement error is a useful modelling pattern for simulation builders. Readers should watch whether the gains hold across other sensors and environments.
PRISM: Learning Realistic Depth via Physics-Grounded Noise Disentanglement with Semantic-Geometric Collaboration
arXiv:2610.29432v1 Announce Type: new Abstract: Real-world physical sensing exhibits complex, heterogeneous noise patterns that deviate significantly from idealized simulation, posing a fundamental bottleneck for sim-to-real transfer. Existing sensor modelings typically treat depth noise as a monolithic black-box process, overlooking the distinct physical mechanisms that govern different error modalities. In this work, we introduce a physics-grounded paradigm that disentangles monolithic noise into two complementary modalities: \textbf{sensing invalidation} and \textbf{measurement inaccuracy}, enabling a tailored treatment of noise sources based on their physical origins. Building on this insight, we propose \textbf{PRISM}, a tripartite framework that distills 3D Visual Foundation Model features as rich spatial-semantic priors for physics-based reasoning. To address the inherent sparsity and class imbalance of invalidation regions, we develop Hierarchical Positive-Prioritized Supervision, integrating multi-scale positive-weighted objectives with a positive-preserving dynamic hard mining strategy to enforce precise artifact delineation. Extensive benchmarks demonstrate that PRISM achieves state-of-the-art fidelity in noisy depth synthesis. Furthermore, downstream robotic experiments show that PRISM facilitates a \textbf{92.1\%} average success rate in the real world, marking a significant improvement over monolithic baselines. Our project page is available at: \resizebox{\linewidth}{!}{\href{https://shawnleung-sii.github.io/prism-icml2026/}{https://shawnleung-sii.github.io/prism-icml2026/}.}
Abstract:Real-world physical sensing exhibits complex, heterogeneous noise patterns that deviate significantly from idealized simulation, posing a fundamental bottleneck for sim-to-real transfer. Existing sensor modelings typically treat depth noise as a monolithic black-box process, overlooking the distinct physical mechanisms that govern different error modalities. In this work, we introduce a physics-grounded paradigm that disentangles monolithic noise into two complementary modalities: \textbf{sensing invalidation} and \textbf{measurement inaccuracy}, enabling a tailored treatment of noise sources based on their physical origins. Building on this insight, we propose \textbf{PRISM}, a tripartite framework that distills 3D Visual Foundation Model features as rich spatial-semantic priors for physics-based reasoning. To address the inherent sparsity and class imbalance of invalidation regions, we develop Hierarchical Positive-Prioritized Supervision, integrating multi-scale positive-weighted objectives with a positive-preserving dynamic hard mining strategy to enforce precise artifact delineation. Extensive benchmarks demonstrate that PRISM achieves state-of-the-art fidelity in noisy depth synthesis. Furthermore, downstream robotic experiments show that PRISM facilitates a \textbf{92.1\%} average success rate in the real world, marking a significant improvement over monolithic baselines. Our project page is available at: \resizebox{\linewidth}{!}{\href{this https URL}{this https URL}.}
Robotics (cs.RO)
arXiv:2610.29432 [cs.RO]
(or arXiv:2610.29432v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.29432 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 272 cuvinte