EntON: Eigenentropy-Optimized Neighborhood Densification in 3D Gaussian Splatting

2026-03-06

AI Summary

This work addresses the limitations of conventional 3D Gaussian splatting, where misalignment between Gaussian centers and object surface geometry degrades reconstruction quality, and existing surface optimization strategies often compromise photometric accuracy. To overcome these issues, we propose a geometry-aware alternating densification strategy that introduces, for the first time, an Eigenentropy metric derived from the eigenvalues of the covariance matrix. This metric, combined with view-space gradients and local geometric structure analyzed via k-nearest-neighbor covariance, adaptively guides the splitting and pruning of Gaussians. Our method significantly enhances geometric detail while preserving high photometric fidelity. Experiments on the DTU and TUM2TWIN datasets demonstrate up to a 33% improvement in geometric accuracy, a 7% gain in rendering quality, a 50% reduction in Gaussian count, and a 23% decrease in training time.

Technology Category

• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

We present a novel Eigenentropy-optimized neighboorhood densification strategy EntON in 3D Gaussian Splatting (3DGS) for geometrically accurate and high-quality rendered 3D reconstruction. While standard 3DGS produces Gaussians whose centers and surfaces are poorly aligned with the underlying object geometry, surface-focused reconstruction methods frequently sacrifice photometric accuracy. In contrast to the conventional densification strategy, which relies on the magnitude of the view-space position gradient, our approach introduces a geometry-aware strategy to guide adaptive splitting and pruning. Specifically, we compute the 3D shape feature Eigenentropy from the eigenvalues of the covariance matrix in the k-nearest neighborhood of each Gaussian center, which quantifies the local structural order. These Eigenentropy values are integrated into an alternating optimization framework: During the optimization process, the algorithm alternates between (i) standard gradient-based densification, which refines regions via view-space gradients, and (ii) Eigenentropy-aware densification, which preferentially densifies Gaussians in low-Eigenentropy (ordered, flat) neighborhoods to better capture fine geometric details on the object surface, and prunes those in high-Eigenentropy (disordered, spherical) regions. We provide quantitative and qualitative evaluations on two benchmark datasets: small-scale DTU dataset and large-scale TUM2TWIN dataset, covering man-made objects and urban scenes. Experiments demonstrate that our Eigenentropy-aware alternating densification strategy improves geometric accuracy by up to 33% and rendering quality by up to 7%, while reducing the number of Gaussians by up to 50% and training time by up to 23%. Overall, EnTON achieves a favorable balance between geometric accuracy, rendering quality and efficiency by avoiding unnecessary scene expansion.

Problem

• 3D Gaussian Splatting
• geometric accuracy
• surface reconstruction
• photometric accuracy
• densification

Innovation

• Eigenentropy
• 3D Gaussian Splatting
• geometry-aware densification
• adaptive splitting and pruning
• surface reconstruction