EVolSplat4D: Efficient Volume-based Gaussian Splatting for 4D Urban Scene Synthesis

TL;DR

EvolSplat4D introduces a unified, feed-forward 4D scene synthesis framework that combines volume-based and pixel-based Gaussian predictions, achieving high-quality, consistent reconstructions of static and dynamic urban scenes efficiently.

Contribution

It presents a novel multi-branch approach that unifies volume-based and pixel-based Gaussian prediction for 4D scene synthesis, surpassing existing methods in accuracy and efficiency.

Findings

Outperforms state-of-the-art methods on multiple datasets.

Achieves superior static and dynamic scene reconstruction accuracy.

Ensures stable 4D reconstruction despite noisy motion priors.

Abstract

Novel view synthesis (NVS) of static and dynamic urban scenes is essential for autonomous driving simulation, yet existing methods often struggle to balance reconstruction time with quality. While state-of-the-art neural radiance fields and 3D Gaussian Splatting approaches achieve photorealism, they often rely on time-consuming per-scene optimization. Conversely, emerging feed-forward methods frequently adopt per-pixel Gaussian representations, which lead to 3D inconsistencies when aggregating multi-view predictions in complex, dynamic environments. We propose EvolSplat4D, a feed-forward framework that moves beyond existing per-pixel paradigms by unifying volume-based and pixel-based Gaussian prediction across three specialized branches. For close-range static regions, we predict consistent geometry of 3D Gaussians over multiple frames directly from a 3D feature volume, complemented by a semantically-enhanced image-based rendering module for predicting their appearance. For dynamic actors, we utilize object-centric canonical spaces and a motion-adjusted rendering module to aggregate temporal features, ensuring stable 4D reconstruction despite noisy motion priors. Far-Field scenery is handled by an efficient per-pixel Gaussian branch to ensure full-scene coverage. Experimental results on the KITTI-360, KITTI, Waymo, and PandaSet datasets show that EvolSplat4D reconstructs both static and dynamic environments with superior accuracy and consistency, outperforming both per-scene optimization and state-of-the-art feed-forward baselines.