Refining Gaussian Splatting: A Volumetric Densification Approach
This work addresses a specific technical bottleneck in 3D reconstruction for computer vision researchers, representing an incremental improvement over existing methods.
The paper tackles the problem of improving novel view synthesis quality in 3D Gaussian Splatting by addressing shortcomings in its densification strategy, introducing a method that uses Gaussian volumes of inertia to guide refinement and testing different point cloud initializations, resulting in performance surpassing 3DGS on the Mip-NeRF 360 dataset.
Achieving high-quality novel view synthesis in 3D Gaussian Splatting (3DGS) often depends on effective point primitive management. The underlying Adaptive Density Control (ADC) process addresses this issue by automating densification and pruning. Yet, the vanilla 3DGS densification strategy shows key shortcomings. To address this issue, in this paper we introduce a novel density control method, which exploits the volumes of inertia associated to each Gaussian function to guide the refinement process. Furthermore, we study the effect of both traditional Structure from Motion (SfM) and Deep Image Matching (DIM) methods for point cloud initialization. Extensive experimental evaluations on the Mip-NeRF 360 dataset demonstrate that our approach surpasses 3DGS in reconstruction quality, delivering encouraging performance across diverse scenes.