EnerGS: Energy-Based Gaussian Splatting with Partial Geometric Priors

2026-04-28

AI Summary

In large-scale outdoor scenes, geometric priors such as LiDAR data are often sparse and unevenly distributed, limiting their effectiveness in enhancing 3D Gaussian splatting reconstruction quality and sometimes even degrading performance. To address this challenge, this work proposes a soft geometric guidance mechanism based on a continuous energy field, which models partially observable geometry as an energy field induced by geometric evidence. This approach guides the optimization of Gaussian primitives without imposing hard constraints, thereby effectively integrating sparse and incomplete geometric priors. The method significantly improves photometric fidelity and geometric stability under both multi-view and monocular settings while mitigating training overfitting.

Abstract

3D Gaussian Splatting (3DGS) has been widely adopted for scene reconstruction, where training inherently constitutes a highly coupled and non-convex optimization problem. Recent works commonly incorporate geometric priors, such as LiDAR measurements, either for initialization or as training constraints, with the goal of improving photometric reconstruction quality. However, in large-scale outdoor scenarios, such geometric supervision is often spatially incomplete and uneven, which limits its effectiveness as a reliable prior and can even be detrimental to the final reconstruction. To address this challenge, we model partially observable geometry as a continuous energy field induced by geometric evidence and propose EnerGS. Rather than enforcing geometry as a hard constraint, EnerGS provides a soft geometric guidance for the optimization of Gaussian primitives, allowing geometric information to steer the optimization process without directly restricting the solution space. Extensive experiments on large-scale outdoor scenes demonstrate that, under both sparse multi-view and monocular settings, EnerGS consistently improves photometric quality and geometric stability, while effectively mitigating overfitting during 3DGS training.

Problem

• 3D Gaussian Splatting
• geometric priors
• incomplete geometry
• scene reconstruction
• outdoor scenes

Innovation

• Energy-Based Modeling
• Gaussian Splatting
• Partial Geometric Priors
• Soft Geometric Guidance
• 3D Scene Reconstruction