What question did this study set out to answer?

This research aims to address the issue of ghostly artifacts in casually captured Neural Radiance Fields (NeRFs).

October 1, 2023

Nerfbusters: Removing Ghostly Artifacts from Casually Captured NeRFs

View Full Paper

Ask AI

Bookmark

Authors

Proposed dataset and 3D diffusion method improve scene quality in casual captures, enhancing evaluation protocols.

This research aims to address the issue of ghostly artifacts in casually captured Neural Radiance Fields (NeRFs).
Proposed a new dataset with two camera trajectories: one for training and one for evaluation.
Developed a 3D diffusion-based method using local 3D priors.
Introduced a density-based score distillation loss for better artifact removal during NeRF optimization.
The proposed method effectively removes floaters and improves scene geometry for casual captures.

Warburg et al. (2023) studied this question.

View Full Paper

Ask AI

Bookmark