3DGBGS: 3D Granular Ball Gaussian Splatting for Compact Novel View Synthesis

2026-07-29

AI Summary

This work addresses the challenge of balancing anchor count, model compactness, and rendering quality in existing anchor-based 3D Gaussian splatting methods when applied to spatially non-uniform sparse SfM point clouds. To this end, the authors introduce granular ball computation, which adaptively partitions the point cloud into three-dimensional granular balls—employing larger balls in smooth regions for compact representation and smaller balls in detailed areas to preserve geometric fidelity. They propose Granular Ball Anchor Initialization (GBAI) and Granular Ball Scale Prior (GBSP) to enable adaptive setting of anchor positions and Gaussian scales. Experiments demonstrate that the method reduces initial anchors by 37.1% and final anchors by 10.0% on average across four benchmarks, decreases model storage by 9.8%, and maintains comparable rendering quality.

Abstract

Three-dimensional Gaussian Splatting (3DGS) enables high-quality real-time novel-view synthesis through explicit Gaussian primitives and differentiable rasterization. 3DGS and Granular Ball Computing (GBC), proposed in 2019, share a natural compatibility in adaptive representation. The efficiency of 3DGS partly stems from a coarse-to-fine and on-demand refinement process that draws on the generation principle of GBC. This connection motivates us to further introduce adaptive granular ball organization into anchor-based 3DGS. Existing anchor-based methods typically construct anchors from sparse SfM point clouds through fixed voxelization, which cannot adequately adapt to spatially non-uniform point distributions and leads to a trade-off among anchor count, model compactness, and rendering quality. To address this issue, we propose 3DGBGS (3D Granular Ball Gaussian Splatting), a compact anchor-based framework for novel-view synthesis. 3DGBGS adaptively partitions SfM point clouds into 3D granular balls, using larger balls to compactly represent smooth and redundant regions and smaller balls to preserve complex geometry and local details. Based on this representation, Granular Ball Anchor Initialization (GBAI) uses granular ball centers to initialize compact anchor positions, while the Granular Ball Scale Prior (GBSP) exploits granular ball radii to provide local scale priors for Gaussian generation. Experiments on four benchmarks show that 3DGBGS reduces initial and final anchors by 37.1% and 10.0%, respectively, and model storage by 9.8% on average, while maintaining comparable rendering quality.

Problem

• 3D Gaussian Splatting
• anchor-based representation
• spatially non-uniform point distributions
• model compactness
• novel view synthesis

Innovation

• Granular Ball Computing
• 3D Gaussian Splatting
• adaptive representation
• anchor initialization
• novel view synthesis