🤖 AI Summary
Existing 3D Gaussian Splatting (3DGS) relies on rasterization, limiting its ability to model physically based lighting effects—such as soft shadows, specular reflections, and global illumination. This work introduces, for the first time, differentiable ray tracing into 3DGS. We propose an efficient ray–ellipsoid intersection algorithm tailored for ellipsoidal Gaussian primitives, enabling explicit lighting modeling and shadow generation. We further design an RGB-confidence ellipsoidal representation and develop a path-tracing–compatible rendering framework that supports hybrid rendering with triangular meshes. Our approach preserves the efficiency of 3DGS in training and real-time rendering while significantly enhancing physical plausibility: PSNR improves by ≥2.1 dB, and high-fidelity soft shadows, accurate specular reflections, and interactive light editing are enabled.
📝 Abstract
3D Gaussian Splatting (3DGS) is a process that enables the direct creation of 3D objects from 2D images. This representation offers numerous advantages, including rapid training and rendering. However, a significant limitation of 3DGS is the challenge of incorporating light and shadow reflections, primarily due to the utilization of rasterization rather than ray tracing for rendering. This paper introduces RaySplats, a model that employs ray-tracing based Gaussian Splatting. Rather than utilizing the projection of Gaussians, our method employs a ray-tracing mechanism, operating directly on Gaussian primitives represented by confidence ellipses with RGB colors. In practice, we compute the intersection between ellipses and rays to construct ray-tracing algorithms, facilitating the incorporation of meshes with Gaussian Splatting models and the addition of lights, shadows, and other related effects.