CVPR2024
Bayes' Rays: Uncertainty Quantification for Neural Radiance Fields
Lily Goli, Cody Reading, Silvia Sellán, Alec Jacobson, Andrea Tagliasacchi
摘要
2 Google DeepMind 3 Adobe Research 4 Train NeRF Novel View Our Spatial Uncertainty Clean Novel View Bayes' Rays thresholding by uncertainty Figure 1. We introduce BayesRays, a post-hoc algorithm to estimate the spatial uncertainty of any pre-trained NeRF of any arbitrary architecture. Our method requires no additional training and can be used to clean up NeRF artifacts caused by occluded or incomplete data.