Stress testing across synthetic depth distributions

Characterize the performance and repair behavior of TileGS on synthetic scenes with sparse, dense, and strongly bimodal near/far depth distributions, including the regimes in which depth-binning overhead dominates and the repair budget becomes active.

Background

The evaluation is restricted to natural real-world captures from three datasets. The paper identifies synthetic scenes with unusually sparse, dense, or bimodal depth distributions as potentially interacting differently with logarithmic depth binning and the selective repair policy.

A systematic stress test remains to be conducted to determine when the auxiliary cost of binning outweighs its benefits and when the repair-entry budget is activated.

References

A systematic stress test over synthetic sparse, dense, and bimodal depth distributions would be useful for characterizing when binning overhead dominates and when the repair budget becomes active; we treat this as future work.

— TileGS: Tile-Local Depth Binning for Gaussian Splatting Rasterization  (2609.03613 - Tan et al., 3 Sep 2026) in Section 6, subsection “Scene coverage”