TileSkipper: Region-Adaptive Tile Pruning for 3D Gaussian Splatting

2026-10-06

AI Summary

This study addresses the inefficiency and content-insensitivity caused by globally uniform truncation thresholds in 3D Gaussian Splatting rasterization. We propose a training-free adaptive truncation strategy that constructs a distortion proxy metric based on front-to-back transmittance and background color. Through offline calibration, byte-level truncation parameters are assigned to each Gaussian, while region-adaptive tile pruning and compiler-level integration further optimize the enumeration process without requiring parameter updates or additional inference overhead. Experimental results demonstrate that our method achieves 1.088× and 1.238× speedups under standard and high-resolution settings, respectively, with negligible PSNR degradation, significantly outperforming the globally calibrated baseline.

Abstract

Tiled 3D Gaussian Splatting rasterizers often use one scene-wide contribution cutoff for tile enumeration, although content differs in its sensitivity to support truncation. TileSkipper selects a static per-Gaussian cutoff policy for a frozen checkpoint. Calibration renders measure candidate pair savings and an isolated-removal distortion proxy that accounts for front transmittance and background color. The method allocates cutoffs across 64 Gaussian groups and accepts policies only after complete renders on disjoint selection views. The exported policy uses one byte per Gaussian, with no parameter updates, additional kernel, or per-frame policy inference. Across 13 scenes from Mip-NeRF 360, Tanks & Temples, and Deep Blending, a fixed-policy AccuTile sweep gives dataset-macro speedups of $1.088\times$ at standard resolution and $1.238\times$ at 3840 pixels wide, with $-0.007/-0.023$ dB mean PSNR change. Six integrations with existing opacity-aware bounds yield $1.009\times$--$1.121\times$ compiler-only speedups. For four ports from $3σ$ rasterizers, we separately attribute the prior exact-bound transition and our incremental gain. Matched-quality ablations show modest gains over scene-global calibration and parity with per-Gaussian control; the standalone comparison with AdaGScale is regime-dependent.

Problem

• 3D Gaussian Splatting
• tile pruning
• rasterization
• rendering acceleration
• contribution cutoff

Innovation

• 3D Gaussian Splatting
• Region-Adaptive Tile Pruning
• Per-Gaussian Cutoff Policy
• Rasterization Acceleration
• Zero-Overhead Inference