Re-Activating Frozen Primitives for 3D Gaussian Splatting

2025-10-21

AI Summary

3D Gaussian Splatting (3D-GS) suffers from over-reconstruction artifacts—such as needle-like distortions and local blurring—in complex scenes. These arise not merely from insufficient splitting, but primarily from gradient dilution and “primitive freezing”: critical-region Gaussians fail to densify effectively, while suboptimally scaled Gaussians become trapped in local optima. To address this, we propose a reactivation mechanism: first, an importance-aware densification criterion precisely identifies regions requiring refinement; second, adaptive parameter perturbation dynamically detects and reactivates frozen primitives. This approach transcends conventional static densification thresholds, enabling targeted optimization of geometrically sensitive areas. Our method significantly suppresses reconstruction artifacts while preserving fine-scale geometry. Evaluated on multiple real-world datasets, it achieves state-of-the-art novel-view synthesis quality and maintains full compatibility with variants such as Pixel-GS.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Sampling/Simulation-based Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

3D Gaussian Splatting (3D-GS) achieves real-time photorealistic novel view synthesis, yet struggles with complex scenes due to over-reconstruction artifacts, manifesting as local blurring and needle-shape distortions. While recent approaches attribute these issues to insufficient splitting of large-scale Gaussians, we identify two fundamental limitations: gradient magnitude dilution during densification and the primitive frozen phenomenon, where essential Gaussian densification is inhibited in complex regions while suboptimally scaled Gaussians become trapped in local optima. To address these challenges, we introduce ReAct-GS, a method founded on the principle of re-activation. Our approach features: (1) an importance-aware densification criterion incorporating $α$-blending weights from multiple viewpoints to re-activate stalled primitive growth in complex regions, and (2) a re-activation mechanism that revitalizes frozen primitives through adaptive parameter perturbations. Comprehensive experiments across diverse real-world datasets demonstrate that ReAct-GS effectively eliminates over-reconstruction artifacts and achieves state-of-the-art performance on standard novel view synthesis metrics while preserving intricate geometric details. Additionally, our re-activation mechanism yields consistent improvements when integrated with other 3D-GS variants such as Pixel-GS, demonstrating its broad applicability.

Problem

• Addresses over-reconstruction artifacts in 3D Gaussian Splatting
• Solves gradient dilution and primitive freezing in complex scenes
• Revitalizes frozen primitives to preserve geometric details

Innovation

• Re-activates frozen primitives via adaptive parameter perturbations
• Uses importance-aware densification with multi-view blending weights
• Eliminates over-reconstruction artifacts while preserving geometric details