ADATEX4D: adaptive texture capacity allocation for 4D gaussian splatting

2026-09-24

AI Summary

This study addresses the storage redundancy in 4D Gaussian Splatting caused by uniform texture resolution across regions with varying detail levels. To this end, we propose an adaptive texture capacity module that introduces a dynamic anisotropic texture allocation mechanism driven by screen-space gradients and local deformation scales. Combined with a packed RGBA tri-plane representation, this approach enables precise, on-demand distribution of texture resources. Experimental evaluations on datasets such as N3DV demonstrate that the proposed method reduces texture storage overhead by over 50%. Furthermore, under a fixed memory budget, it significantly enhances dynamic scene reconstruction quality while lowering peak GPU memory consumption.

Abstract

Textured Gaussians improve local appearance capacity, but assigning the same texture resolution to every primitive wastes storage on low-detail or weakly visible regions. We introduce AdaTex4D, an adaptive texture-capacity module for deformation-based 4D Gaussian Splatting. Each Gaussian carries packed RGBA triplanes whose two axes grow independently according to visibility normalized screen-space gradients and deformed local scales. Experiments on N3DV and PanopticSports show that AdaTex4D reduces texture storage by more than half while preserving reconstruction quality. Under fixed memory budgets, adaptive allocation also improves quality over uniform texture assignment and reduces overall model and peak memory. These results show that dynamic, anisotropic texture allocation provides a more efficient way to distribute local appearance capacity in 4D Gaussian representations.

Problem

• 4D Gaussian Splatting
• texture capacity allocation
• storage efficiency
• dynamic scene representation

Innovation

• 4D Gaussian Splatting
• Adaptive Texture Allocation
• Anisotropic Triplanes
• Memory Efficiency
• Dynamic Rendering