Efficient Confidence-Aware Densification for 3DGS

Key points

• Because opacity stays high where the scene is already well reconstructed, this sampling sends most new primitives to the foreground and leaves the background under-reconstructed.1
• At low budgets, prior-based methods such as ConeGS and EDGS outperform it, at the cost of a pre-trained radiance field or dense image matches and a separate initialization stage.1
• Our confidence-aware densification keeps the fixed budget and gradual growth of 3DGS-MCMC, but places primitives by the image error instead of the opacity.1
• While inherently prior-free, our densification can also exploit priors: from the ConeGS or EDGS initialization, it matches ConeGS in PSNR and SSIM at 100k and is best in all three metrics at 500k.1

Sources (1)

• [1]Efficient Confidence-Aware Densification for 3DGS
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 12:59 PM

  3DGS-MCMC has replaced the gradient-driven density control of 3D Gaussian Splatting with opacity-based sampling and relocation, and is now the de facto standard for prior-free densification under a fixed primitive budget.

  Because opacity stays high where the scene is already well reconstructed, this sampling sends most new primitives to the foreground and leaves the background under-reconstructed.

Extractive summary: sentences quoted from the sources.