Modeling Ambient Scene Dynamics for Free-view Synthesis

TL;DR

This paper presents a new method for dynamic free-view synthesis of ambient scenes from monocular captures, enhancing realism and generalization over previous approaches that were limited to static or bounded scenes.

Contribution

It introduces a novel approach leveraging ambient motion periodicity and regularization to extend 3D Gaussian Splatting for dynamic scene reconstruction from monocular data.

Findings

Achieves high-quality photorealistic novel view synthesis of natural scenes.

Improves memory efficiency and visual quality of 3D scene reconstructions.

Demonstrates generalization to unseen motions in ambient scene synthesis.

Abstract

We introduce a novel method for dynamic free-view synthesis of an ambient scenes from a monocular capture bringing a immersive quality to the viewing experience. Our method builds upon the recent advancements in 3D Gaussian Splatting (3DGS) that can faithfully reconstruct complex static scenes. Previous attempts to extend 3DGS to represent dynamics have been confined to bounded scenes or require multi-camera captures, and often fail to generalize to unseen motions, limiting their practical application. Our approach overcomes these constraints by leveraging the periodicity of ambient motions to learn the motion trajectory model, coupled with careful regularization. We also propose important practical strategies to improve the visual quality of the baseline 3DGS static reconstructions and to improve memory efficiency critical for GPU-memory intensive learning. We demonstrate high-quality photorealistic novel view synthesis of several ambient natural scenes with intricate textures and fine structural elements.