RobustNeRF: Ignoring Distractors with Robust Losses
Citations Over TimeTop 10% of 2023 papers
Abstract
Neural radiance fields (NeRF) excel at synthesizing new views given multi-view, calibrated images of a static scene. When scenes include distractors, which are not persistent during image capture (moving objects, lighting variations, shadows), artifacts appear as view-dependent effects or ‘floaters’. To cope with distractors, we advocate a form of robust estimation for NeRF training, modeling distractors in training data as outliers of an optimization problem. Our method successfully removes outliers from a scene and improves upon our baselines, on synthetic and real-world scenes. Our technique is simple to incorporate in modern NeRF frameworks, with few hyper-parameters. It does not assume a priori knowledge of the types of distractors, and is instead focused on the optimization problem rather than pre-processing or modeling transient objects. More results at https://robustnerf.github.io/public.
Related Papers
- → Use of the radiance distribution to measure the optical absorption coefficient in the ocean(1989)49 cited
- → ON THE INSTRUMENT FOR MEASURING ANGULAR DISTRIBUTION OF UNDERWATER RADIANCE(1962)16 cited
- → A guide to performing hazard evaluation using radiance and integrated radiance(2005)1 cited
- Research on the Influence of Solar Radiance,Sky Radiance and Sea Surface Radiance to the Efficiency of Ocean Total Radiance(2012)
- → Calculation of infrared radiance fluctuations using the Strategic High Altitude Radiance Code, SHARC, for space-based remote sensing applications(1995)