MrNeRF (@janusch_patas)
2024-11-12 | ❤️ 92 | 🔁 8
GaussianSpa: An “Optimizing-Sparsifying” Simplification Framework for Compact and High-Quality 3D Gaussian Splatting
TL;DR: High quality compaction! Damn, that looks impressive!
Contributions (cited): • We propose a general 3DGS simplification framework that formulates the simplification objective as an optimization problem and solves it in the 3DGS training process.
In solving the formulated optimization problem, our proposed framework gradually restricts Gaussians to the target sparsity constraint without explicitly removing a specific number of points. Hence, GaussianSpa can maximally maintain and smoothly transfer the information to the sparse Gaussians from the original model.
• We propose an efficient “optimizing-sparsifying” solution for the formulated problem, which can be integrated into the 3DGS training with negligible costs, separately solving two sub-problems.
In the “optimizing” step, we optimize the original loss function attached by a regularization with gradient descent. In the “sparsifying” step, we analytically project the auxiliary Gaussians onto the constrained sparse space.
• We comprehensively evaluate GaussianSpa through extensive experiments on various complex scenes, demonstrating improved rendering quality compared to existing approaches.
Particularly, with as high as 10× fewer number of Gaussians than the vanilla 3DGS, GaussianSpa achieves an average 0.4 dB improvement on the Mip-NeRF 360 [4] and Tanks&Temples [27] datasets, 0.9 dB on Deep Blending [23] dataset.
Furthermore, we conduct various visual quality assessments, showing that GaussianSpa exhibits high-quality rendering of details and sparse 3D Gaussian views.
미디어
Tags
domain-vision-3d domain-rendering domain-ai-ml domain-dev-tools