DynaSplat: Dynamic-Static Gaussian Splatting with Hierarchical Motion Decomposition for Scene Reconstruction

TL;DR

DynaSplat advances dynamic scene reconstruction by combining static-dynamic separation, hierarchical motion modeling, and physically-based opacity estimation, resulting in more accurate, realistic, and efficient reconstructions of complex, non-rigid environments.

Contribution

The paper introduces DynaSplat, a novel method that extends Gaussian Splatting with dynamic-static separation and hierarchical motion modeling for improved scene reconstruction.

Findings

Outperforms state-of-the-art methods in accuracy and realism

Effectively handles complex non-rigid motions

Provides a more compact and efficient reconstruction process

Abstract

Reconstructing intricate, ever-changing environments remains a central ambition in computer vision, yet existing solutions often crumble before the complexity of real-world dynamics. We present DynaSplat, an approach that extends Gaussian Splatting to dynamic scenes by integrating dynamic-static separation and hierarchical motion modeling. First, we classify scene elements as static or dynamic through a novel fusion of deformation offset statistics and 2D motion flow consistency, refining our spatial representation to focus precisely where motion matters. We then introduce a hierarchical motion modeling strategy that captures both coarse global transformations and fine-grained local movements, enabling accurate handling of intricate, non-rigid motions. Finally, we integrate physically-based opacity estimation to ensure visually coherent reconstructions, even under challenging occlusions and perspective shifts. Extensive experiments on challenging datasets reveal that DynaSplat not only surpasses state-of-the-art alternatives in accuracy and realism but also provides a more intuitive, compact, and efficient route to dynamic scene reconstruction.