ParetoGS improves 3D Gaussian splatting quality across budgets
ParetoGS: High-Quality Continuous-Budget 3D Gaussian Splatting with Sparse Adapted Anchors
Computer Vision and Pattern Recognition
Summary
3D Gaussian splatting creates detailed digital scenes using many small shapes called Gaussians, but fewer Gaussians usually means lower quality. The authors propose ParetoGS, a method that adjusts these shapes across different amounts of Gaussians without redoing everything each time. ParetoGS keeps track of important shapes and smoothly changes their details to maintain high quality even when fewer shapes are used. This makes rendering flexible and efficient, adapting well to different levels of detail.
What this means in practice
- •For augmented reality developers: Generate high-quality 3D scenes with adjustable detail levels to optimize performance on devices with varying processing power.
- •For game engine teams: Create scalable 3D assets that maintain visual quality at different performance budgets without retraining models for each target.
Authors
Xiaobin Deng, Changyu Diao, Min Li, Duanqing Xu
Abstract
Scalable 3D Gaussian Splatting allows one scene representation to serve different Gaussian-count budgets, yet its quality-budget curve may lag behind models optimized separately for each budget. Reducing the number of Gaussians changes both the selected members and the reconstruction tasks carried by the survivors; parameters inherited from a high-budget model may fit a low-budget representation poorly. We propose ParetoGS, which describes member changes and parameter evolution with a small set of adapted budget anchors and decodes intermediate budgets without separate optimization at every requested count. ParetoGS alternates pruning and fixed-member adaptation to build nested anchors with persistent identities, and encodes their parameter changes relative to the full model in a compact residual atlas. At decoding, it interpolates the states of shared members and gradually introduces the rendering contributions of new members, connecting sparse anchors into a continuous-budget representation. Across 13 scenes, ParetoGS attains better average PSNR, SSIM, and LPIPS with fewer Gaussians than the compared prefix methods at 10% of each method's own full model. From a shared full model, it also outperforms FlexGS-imp at its trained 1%, 5%, and 10% budgets and NanoGS-imp at the 1% and 0.1% budgets. On Garden, its non-anchor quality approaches an empirical reference from independently optimized GNS models at higher retention ratios and exceeds it at some lower-budget points. Across 13 scenes, encoding reduces mean additional state from 390.0 to 85.7 MB with an average PSNR change of 0.01 dB. Project page: https://xiaobin2001.github.io/ParetoGS-Web/