ExploreGS: Explorable 3D Scene Reconstruction with Virtual Camera Samplings and Diffusion Priors

TL;DR

ExploreGS introduces a novel pipeline combining virtual camera sampling and diffusion priors to improve 3D scene reconstruction, enabling artifact-free rendering from arbitrary viewpoints and advancing scene exploration capabilities.

Contribution

The paper presents a new method integrating virtual camera placement and diffusion priors to enhance 3D Gaussian Splatting reconstructions, addressing artifacts and missing regions.

Findings

Outperforms existing 3DGS methods in quality and coverage

Enables seamless scene exploration from arbitrary viewpoints

Introduces Wild-Explore benchmark for challenging scene evaluation

Abstract

Recent advances in novel view synthesis (NVS) have enabled real-time rendering with 3D Gaussian Splatting (3DGS). However, existing methods struggle with artifacts and missing regions when rendering from viewpoints that deviate from the training trajectory, limiting seamless scene exploration. To address this, we propose a 3DGS-based pipeline that generates additional training views to enhance reconstruction. We introduce an information-gain-driven virtual camera placement strategy to maximize scene coverage, followed by video diffusion priors to refine rendered results. Fine-tuning 3D Gaussians with these enhanced views significantly improves reconstruction quality. To evaluate our method, we present Wild-Explore, a benchmark designed for challenging scene exploration. Experiments demonstrate that our approach outperforms existing 3DGS-based methods, enabling high-quality, artifact-free rendering from arbitrary viewpoints. https://exploregs.github.io