Gaussian Set Surface Reconstruction through Per-Gaussian Optimization
This addresses geometry reconstruction issues in 3D Gaussian Splatting for 3D scene editing and environment generation, representing an incremental improvement over existing methods.
The paper tackles the problem of inaccurate scene geometry reconstruction in 3D Gaussian Splatting by proposing Gaussian Set Surface Reconstruction (GSSR), which optimizes Gaussian placement to align with latent surfaces and improves geometric precision while maintaining rendering quality.
3D Gaussian Splatting (3DGS) effectively synthesizes novel views through its flexible representation, yet fails to accurately reconstruct scene geometry. While modern variants like PGSR introduce additional losses to ensure proper depth and normal maps through Gaussian fusion, they still neglect individual placement optimization. This results in unevenly distributed Gaussians that deviate from the latent surface, complicating both reconstruction refinement and scene editing. Motivated by pioneering work on Point Set Surfaces, we propose Gaussian Set Surface Reconstruction (GSSR), a method designed to distribute Gaussians evenly along the latent surface while aligning their dominant normals with the surface normal. GSSR enforces fine-grained geometric alignment through a combination of pixel-level and Gaussian-level single-view normal consistency and multi-view photometric consistency, optimizing both local and global perspectives. To further refine the representation, we introduce an opacity regularization loss to eliminate redundant Gaussians and apply periodic depth- and normal-guided Gaussian reinitialization for a cleaner, more uniform spatial distribution. Our reconstruction results demonstrate significantly improved geometric precision in Gaussian placement, enabling intuitive scene editing and efficient generation of novel Gaussian-based 3D environments. Extensive experiments validate GSSR's effectiveness, showing enhanced geometric accuracy while preserving high-quality rendering performance.