ConFixGS: Learning to Fix Feedforward 3D Gaussian Splatting with Confidence-Aware Diffusion Priors in Driving Scenes

TL;DR

ConFixGS is a novel method that enhances feedforward 3D Gaussian Splatting in driving scenes by integrating confidence-aware diffusion priors for improved view synthesis.

Contribution

It introduces a confidence-aware diffusion prior framework that refines feedforward 3DGS using local pseudo-targets and cross-view validation.

Findings

PSNR improved by up to 3.68 dB on benchmarks

FID score reduced by nearly 50%

Enhanced robustness in challenging novel view synthesis scenarios

Abstract

Feedforward 3D Gaussian Splatting (3DGS) often struggles in trajectory-based sparse-view driving scenes. Existing Gaussian repair methods mainly target optimization-based 3DGS, while diffusion-based repair is typically restricted to iterative refinement near observed viewpoints, leaving feedforward 3DGS repair underexplored. We propose ConFixGS, a plug-and-play method that learns to fix feedforward 3DGS with confidence-aware diffusion priors. Starting from a pretrained feedforward model, ConFixGS generates diffusion-enhanced local pseudo-targets and validates them through reprojection-based cross-checking against support views. The resulting dense confidence maps guide refinement, enhancing reliable details while suppressing hallucinated or inconsistent evidence. On Waymo, nuScenes, and KITTI, ConFixGS improves challenging novel view synthesis, with PSNR gains of up to 3.68 dB and FID reduced by nearly half. Our results highlight confidence-aware fusion of generative priors and support-view consistency as a key principle for robust feedforward 3D driving scene reconstruction.