🤖 AI Summary
Existing Gaussian splatting-based rendering methods rely on proxy geometry for ray tracing, resulting in complex intermediate mesh construction, expensive ray-triangle intersection computations, and difficulty unifying differentiable ray tracing and rasterization. This paper introduces the first rendering pipeline that adopts triangles as unified, differentiable primitives—eliminating proxy geometry entirely and enabling direct differentiable ray tracing and rasterization over optimized Triangle Splatting representations. Our core innovation is elevating triangles to end-to-end trainable, general-purpose rendering primitives, thereby achieving deep integration of both rendering paradigms. The method maintains real-time performance while significantly improving novel-view synthesis quality—outperforming state-of-the-art Gaussian-based ray tracing approaches—and enables immediate rasterized rendering of optimized results.
📝 Abstract
Ray tracing 3D Gaussian particles enables realistic effects such as depth of field, refractions, and flexible camera modeling for novel-view synthesis. However, existing methods trace Gaussians through proxy geometry, which requires constructing complex intermediate meshes and performing costly intersection tests. This limitation arises because Gaussian-based particles are not well suited as unified primitives for both ray tracing and rasterization. In this work, we propose a differentiable triangle-based ray tracing pipeline that directly treats triangles as rendering primitives without relying on any proxy geometry. Our results show that the proposed method achieves significantly higher rendering quality than existing ray tracing approaches while maintaining real-time rendering performance. Moreover, our pipeline can directly render triangles optimized by the rasterization-based method Triangle Splatting, thus unifying the primitives used in novel-view synthesis.