Search papers, labs, and topics across Lattice.
This paper introduces 3D Gaussian Accelerated Ray Tracing (3DGART), a novel training framework that optimizes the backward propagation process in ray-traced Gaussian rendering by reorganizing it around primitives instead of pixels. This approach significantly reduces atomic contention and thread serialization during gradient accumulation, leading to a training speedup of approximately 3-3.5 times compared to traditional per-pixel methods and about 4 times faster than existing Gaussian ray tracing techniques. The results demonstrate that 3DGART not only enhances training efficiency but also maintains high rendering quality, making ray-traced Gaussian training feasible for real-time applications.
Reorganizing backward propagation around primitives slashes training time for ray-traced Gaussian rendering by up to 4x while improving output quality.
3D Gaussian Splatting has made Gaussian primitives a highly efficient representation for real-time novel view synthesis, but its rasterisation-based formulation relies on screen-space approximations that limit accurate view-dependent ordering and the integration of secondary ray effects such as reflections, refractions, and shadows. Gaussian ray tracing addresses these limitations by evaluating explicit ray-primitive intersections, yet it remains costly to train. We observe that the main bottleneck is not ray traversal alone, but the pixel-centric backward propagation, where many threads concurrently accumulate gradients into the same primitive parameters, causing severe atomic contention and thread serialisation. We present 3DGART, a practical training framework for ray-traced Gaussian rendering. Our key idea is to reorganise backward propagation around primitives rather than pixels. Using conservative perspective-correct screen-space bounds, we build a compact intermediate buffer and a tile-primitive mapping that allows each thread to accumulate the contribution of one primitive over its covered pixels within a tile. This transforms gradient computation from a contention-heavy scatter operation into a structured gather-like process. On Mip-NeRF 360, 3DGART achieves an $\approx 3-3.5\times$ raw training speedup over per-pixel baseline and $\approx4 \times$ over 3DGRT on Mip-NeRF 360 while improving quality. More importantly, 3DGART makes fully ray-traced Gaussian training practical, reaching runtimes competitive with rasterisation-based pipelines while preserving benefits of ray tracing.
