ExtraGS: Geometric-Aware Trajectory Extrapolation with Uncertainty-Guided Generative Priors

TL;DR

ExtraGS is a novel framework that improves the realism and geometric consistency of extrapolated driving views by integrating geometric and generative priors with uncertainty estimation, addressing limitations of previous methods.

Contribution

It introduces a new hybrid Gaussian-SDF based Road Surface Gaussian representation, Far Field Gaussians, and a self-supervised uncertainty estimation framework for trajectory extrapolation.

Findings

Enhanced geometric consistency in extrapolated views

Improved realism and fidelity in generated scenes

Effective handling of distant objects with learnable scaling

Abstract

Synthesizing extrapolated views from recorded driving logs is critical for simulating driving scenes for autonomous driving vehicles, yet it remains a challenging task. Recent methods leverage generative priors as pseudo ground truth, but often lead to poor geometric consistency and over-smoothed renderings. To address these limitations, we propose ExtraGS, a holistic framework for trajectory extrapolation that integrates both geometric and generative priors. At the core of ExtraGS is a novel Road Surface Gaussian(RSG) representation based on a hybrid Gaussian-Signed Distance Function (SDF) design, and Far Field Gaussians (FFG) that use learnable scaling factors to efficiently handle distant objects. Furthermore, we develop a self-supervised uncertainty estimation framework based on spherical harmonics that enables selective integration of generative priors only where extrapolation artifacts occur. Extensive experiments on multiple datasets, diverse multi-camera setups, and various generative priors demonstrate that ExtraGS significantly enhances the realism and geometric consistency of extrapolated views, while preserving high fidelity along the original trajectory.