AnchorSplat: Fast and Structure Consistent Detail Synthesis for Gaussian Splatting

2026-07-01

AI Summary

This work addresses the lack of fine geometric detail and texture noise in 3D Gaussian Splatting (3DGS) assets, problems exacerbated by existing 2D-based enhancement methods that often introduce multi-view inconsistencies and incur high computational costs. To overcome these limitations, we propose AnchorSplat—the first end-to-end, 3D-native deep network capable of directly enhancing 3D structural details without requiring the original multi-view images. Our approach introduces a point-anchoring mechanism to ensure geometric consistency, replaces iterative optimization with a single multiplicative densification step, and achieves strict source-free, zero-shot generalization. Evaluated on our newly curated large-scale benchmark, 3DGS-SR, AnchorSplat attains state-of-the-art performance, offers inference speeds up to 10⁵ times faster than conventional optimization-based methods, and demonstrates strong generalization across both generative model outputs and real-world scanned data.

Abstract

3D Gaussian Splatting (3DGS) has emerged as a powerful representation for high-fidelity rendering. However, existing assets often suffer from quality bottlenecks such as missing details and texture noise. Prior attempts to enhance these assets via 2D image processing introduce multi-view inconsistencies and high computational costs. In this paper, we propose a novel 3D-native refinement paradigm named AnchorSplat. AnchorSplat is an end-to-end deep network operating directly on 3D structures, avoiding the expensive optimization overhead of traditional 3D-2D-3D pipelines. Crucially, AnchorSplat is a strictly source-free solution requiring no original multi-view images. Central to the proposed method is the Point Anchor Mechanism, which enforces geometric consistency via local offset constraints, mitigating ill-posed mapping and gradient confounding. Furthermore, AnchorSplat replaces iterative densification with a single-pass multiplication mechanism. To facilitate research, we construct 3DGS-SR, the first large-scale benchmark for this task. Experiments demonstrate state-of-the-art results on the 3DGS-SR dataset, with throughput up to $10^5$ times faster than optimization methods. Notably, AnchorSplat exhibits robust zero-shot generalization across diverse data distributions, including generative model outputs and real-world scans.

Problem

• 3D Gaussian Splatting
• detail synthesis
• multi-view consistency
• texture noise
• quality enhancement

Innovation

• AnchorSplat
• 3D Gaussian Splatting
• Point Anchor Mechanism
• source-free refinement
• zero-shot generalization