May 1, 2024Open Access

Efficient Data-driven Scene Simulation using Robotic Surgery Videos via Physics-embedded 3D Gaussians

View Full Paper

Ask AI

Bookmark

Stereo endoscopic video reconstruction completes in a few minutes, and physics-embedded 3D Gaussians generate plausible soft body simulations at speeds approaching real-time.
Simulation framework integrates depth supervision with anisotropy regularization, and applies the material point method to 3D Gaussians for realistic scene deformations.
May enable scalable data-driven scene modeling across robotic surgery datasets, highlighting substantial potential to enhance simulation variety for education and robot learning.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Simulation study demonstrates efficient 3D scene reconstruction and soft tissue deformation in robotic surgery video datasets, highlighting broad utility for surgical education.

Cite This Study

Yang et al. (2024) studied this question.

synapsesocial.com/papers/68e6c5cfb6db6435876441d4 https://doi.org/10.48550/arxiv.2405.00956

View Full Paper

Ask AI

Bookmark

Implication

Simulation study demonstrates efficient 3D scene reconstruction and soft tissue deformation in robotic surgery video datasets, highlighting broad utility for surgical education.

Efficient Data-driven Scene Simulation using Robotic Surgery Videos via Physics-embedded 3D Gaussians

Key Points

Discussion

Implication

Cite This Study

Efficient Data-driven Scene Simulation using Robotic Surgery Videos via Physics-embedded 3D Gaussians

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider