SPARE-GS: Structural Parsimony and Resource Efficiency for 3D Gaussian Splatting

2026-07-17

AI Summary

This work addresses the inefficiency and lack of compactness in 3D Gaussian Splatting caused by redundant primitives. It formulates the structural evolution of the representation as a global budget-constrained optimization problem and introduces a plug-and-play framework that dynamically coordinates densification, pruning, and adaptive termination based on the principle of marginal utility balancing. The framework achieves balanced primitive allocation through capacity-normalized regional demand estimation, adaptive target quota assignment, and a budget-deviation-driven regulation mechanism. Experiments demonstrate that the proposed method reduces the number of primitives by 30.38% and training time by 23.81% on average, while improving PSNR and significantly enhancing the rate-distortion performance of downstream compression and pruning techniques.

Abstract

3D Gaussian Splatting (3DGS) achieves high-fidelity novel view synthesis in real-time; however its training efficiency and representation compactness are hindered by excessive primitive proliferation. To address this challenge, we formulate the structural evolution of 3DGS as a global budget-constrained optimization problem and derive an optimality condition, which requires the marginal utility of structural resources to be balanced across spatial regions under a finite primitive budget. Based on this formulation, we propose SPARE-GS, a general plug-and-play framework that dynamically aligns the distribution of 3D Gaussian primitives with regional representational demand. SPARE-GS estimates capacity-normalized regional demand, assigns adaptive target quotas, and uses regional budget deviations to coordinate densification, pruning and adaptive termination toward a more balanced structural allocation. Extensive experiments across standard, accelerated, and structure-enhanced 3DGS pipelines demonstrate that SPARE-GS reduces the Gaussian count and training time by an average of 30.38% and 23.81%, respectively, while improving the average PSNR. Moreover, the resulting compact representations reduce downstream processing time and improve the rate-distortion performance of diverse compression and pruning methods, demonstrating the broad applicability of global structural budget regulation.

Problem

• 3D Gaussian Splatting
• structural parsimony
• resource efficiency
• primitive proliferation
• representation compactness

Innovation

• Structural Parsimony
• Resource Efficiency
• 3D Gaussian Splatting
• Budget-Constrained Optimization
• Adaptive Pruning