BudSplat improves 3D view rendering under a limited budget
BudSplat: Feed-forward 3D Gaussian Splatting under a Rendering Budget
Computer Vision and Pattern Recognition
Summary
Creating 3D images from lots of photos can get slow because computers have to handle many little pieces called Gaussians. The authors found that many pieces are often repetitive and wasteful, making rendering inefficient. They developed BudSplat, a method that smartly picks the most important pieces to keep the 3D image detailed while staying fast. This approach balances good picture quality with quicker rendering times.
What this means in practice
- •For game developers: Render complex 3D scenes efficiently with limited hardware resources by reducing redundant 3D primitives while preserving visual detail.
- •For augmented reality engineers: Deliver real-time 3D view synthesis on mobile devices by pruning less important primitives to meet rendering budgets without sacrificing quality.
Authors
Muyu Xu, Fangneng Zhan, Yu Wei, Ling Shao, Shijian Lu
Abstract
Feed-forward 3D Gaussian Splatting has recently emerged as a promising alternative to per-scene optimization, enabling fast novel view synthesis by predicting Gaussian primitives directly from input images. However, it quickly faces a bottleneck of rendering efficiency while scaling from sparse to dense multi-view inputs, which offer more geometric constraints but produce a rapidly growing set of candidate Gaussians. Existing solutions cannot reliably distinguish redundant primitives, leading to many redundant primitives from repeatedly observed regions and increased rendering cost without much gain in visual quality. We propose BudSplat, a compact feed-forward 3DGS framework that learns to keep the most valuable Gaussians under a limited primitive count. BudSplat has two novel designs. The first is a detail-aware importance score that is extracted from multi-level visual tokens, providing an explicit cue for regions where pruning is likely to harm perceptual quality. The score can be injected into voxel fusion to preserve high-value local contributors during multi-view aggregation. The second is quality-aware Gaussian pruning that selects a compact set of primitives by jointly considering rendering contribution and detail importance. The quality-aware pruning enables local quota redistribution that helps avoid over-allocating primitives to large redundant regions. Experiments on RE10K, DL3DV, and ACID show that BudSplat achieves a favorable trade-off among rendering quality and efficiency.