Gaussian Opacity Fields: Efficient Adaptive Surface Reconstruction in Unbounded Scenes
This work addresses surface reconstruction for unbounded scenes in computer vision, offering an efficient method that outperforms existing 3DGS-based approaches and competes with neural implicit methods, though it is incremental as it builds on 3D Gaussian Splatting.
The paper tackles the challenge of surface reconstruction from 3D Gaussian Splatting by introducing Gaussian Opacity Fields, which directly extract geometry from 3D Gaussians without needing Poisson reconstruction or TSDF fusion, resulting in improved surface quality and competitive performance in novel view synthesis.
Recently, 3D Gaussian Splatting (3DGS) has demonstrated impressive novel view synthesis results, while allowing the rendering of high-resolution images in real-time. However, leveraging 3D Gaussians for surface reconstruction poses significant challenges due to the explicit and disconnected nature of 3D Gaussians. In this work, we present Gaussian Opacity Fields (GOF), a novel approach for efficient, high-quality, and adaptive surface reconstruction in unbounded scenes. Our GOF is derived from ray-tracing-based volume rendering of 3D Gaussians, enabling direct geometry extraction from 3D Gaussians by identifying its levelset, without resorting to Poisson reconstruction or TSDF fusion as in previous work. We approximate the surface normal of Gaussians as the normal of the ray-Gaussian intersection plane, enabling the application of regularization that significantly enhances geometry. Furthermore, we develop an efficient geometry extraction method utilizing Marching Tetrahedra, where the tetrahedral grids are induced from 3D Gaussians and thus adapt to the scene's complexity. Our evaluations reveal that GOF surpasses existing 3DGS-based methods in surface reconstruction and novel view synthesis. Further, it compares favorably to or even outperforms, neural implicit methods in both quality and speed.