Relaxed Rigidity with Ray-based Grouping for Dynamic Gaussian Splatting

2026-03-25

AI Summary

Existing dynamic 3D Gaussian reconstruction methods often suffer from local geometric distortions in monocular videos due to physically implausible Gaussian motion. To address this, we propose a ray-driven Gaussian grouping mechanism that requires no external priors such as optical flow or 2D trajectories. Our approach clusters Gaussians intersected by the same view ray whose alpha-compositing weights exceed a threshold and enforces spatiotemporal geometric consistency constraints to stabilize local structures and enable more physically plausible dynamic modeling. This method significantly improves temporal coherence and reconstruction quality, outperforming current approaches on multiple challenging monocular datasets. Furthermore, it has been successfully integrated into two distinct baseline models, demonstrating its generality and effectiveness.

Technology Category

• Computer Vision: Motion & Tracking
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Machine Learning: Learning with Manifolds

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

The reconstruction of dynamic 3D scenes using 3D Gaussian Splatting has shown significant promise. A key challenge, however, remains in modeling realistic motion, as most methods fail to align the motion of Gaussians with real-world physical dynamics. This misalignment is particularly problematic for monocular video datasets, where failing to maintain coherent motion undermines local geometric structure, ultimately leading to degraded reconstruction quality. Consequently, many state-of-the-art approaches rely heavily on external priors, such as optical flow or 2D tracks, to enforce temporal coherence. In this work, we propose a novel method to explicitly preserve the local geometric structure of Gaussians across time in 4D scenes. Our core idea is to introduce a view-space ray grouping strategy that clusters Gaussians intersected by the same ray, considering only those whose $α$-blending weights exceed a threshold. We then apply constraints to these groups to maintain a consistent spatial distribution, effectively preserving their local geometry. This approach enforces a more physically plausible motion model by ensuring that local geometry remains stable over time, eliminating the reliance on external guidance. We demonstrate the efficacy of our method by integrating it into two distinct baseline models. Extensive experiments on challenging monocular datasets show that our approach significantly outperforms existing methods, achieving superior temporal consistency and reconstruction quality.

Problem

• Dynamic Gaussian Splatting
• motion modeling
• temporal coherence
• monocular video
• geometric consistency

Innovation

• Ray-based Grouping
• Dynamic Gaussian Splatting
• Local Geometric Consistency
• Temporal Coherence
• View-space Clustering