Anchor guided camera placement speeds up 3d model building
AGILE-GS: Anchor-Guided Fast Next-Best-View Selection for Active 3D Gaussian Splatting
Computer Vision and Pattern RecognitionRobotics
Summary
Building detailed 3D models usually requires taking many pictures from different spots, and choosing the best spots is important but slow. The authors developed a new method that first finds an ideal viewpoint where the model is most uncertain, then quickly picks candidate camera positions near that viewpoint without costly calculations. This approach either picks the closest candidate immediately or ranks a small shortlist with detailed info, making the selection process much faster. Their tests show their method matches or surpasses existing approaches but runs 10 to 100 times faster.
What this means in practice
- •For robotics teams: Enable robots to quickly choose optimal camera positions for 3D mapping, reducing downtime during exploration.
- •For augmented reality developers: Accelerate 3D scene reconstruction by selecting informative views much faster, improving real-time AR experiences.
Authors
Amirhossein Mollaei Khass, Nader Motee
Abstract
Radiance fields need hundreds of views, and their placement matters as much as their number. Next-best-view (NBV) selection for 3D Gaussian Splatting (3DGS) usually scores every candidate in the pool and keeps one. Searching for information and choosing a camera, however, are separable problems. We present AGILE-GS, an anchor-guided NBV method that separates the two. A virtual anchor pose is optimized on SE(3) by Riemannian gradient ascent on expected information gain. It need not be reachable or in the pool; it marks where the model is most uncertain. Candidates are scored against the anchor's viewing geometry, and a greedy ridge-leverage step distills the pool into a small, non-redundant shortlist without rendering any candidate. The shortlist can be used in two ways. AGILE-GS takes the first view on it as the next view, so no Fisher information is computed for any candidate. AGILE-GS+ computes the Fisher information gain of each shortlisted view and picks the best, so the expensive evaluation runs on a handful of views rather than the whole pool. On standard benchmarks and in closed-loop embodied acquisition, both match or exceed existing baselines while cutting selection latency by one to two orders of magnitude.