English

R5DGS: Semantic-Aware 4D Gaussian Splatting with Rigid Body Constraints for Efficient Dynamic Scene Reconstruction

Computer Vision and Pattern Recognition 2026-05-26 v1

Abstract

Reconstructing and predicting dynamic 3D scenes from multi-view videos is a foundational task for robotics, AR/VR, and digital twins. Recent physics-informed Gaussian Splatting methods achieve impressive future frame extrapolation but lack semantic awareness and suffer from large computational overhead. We introduce R5DGS\textbf{R5DGS}, a framework that augments a physics-driven 4D Gaussian representation with compact Identity Encoding vectors, enabling precise Gaussian-to-object association. By constructing an offline CLIP-based object lookup table, we support open-vocabulary text prompting to retrieve and render object-specific Gaussians across arbitrary timestamps and viewpoints. Furthermore, we propose a rigid-body inference constraint that predicts and integrates physical dynamics exclusively for object centroids, propagating motion to associated Gaussians via relative transformations. This optimization yields a 11 FPS speedup during extrapolation without compromising trajectories plausibility.

Keywords

Cite

@article{arxiv.2605.25909,
  title  = {R5DGS: Semantic-Aware 4D Gaussian Splatting with Rigid Body Constraints for Efficient Dynamic Scene Reconstruction},
  author = {Denis Gridusov and Maxim Popov and Sergey Kolyubin},
  journal= {arXiv preprint arXiv:2605.25909},
  year   = {2026}
}

Comments

Code: https://github.com/be2rlab/r5dgs