Floaters No More: Radiance Field Gradient Scaling for Improved Near-Camera Training

TL;DR

This paper introduces a simple gradient scaling method for NeRF training that addresses background collapse caused by sampling imbalance near cameras, eliminating the need for near planes and improving scene reconstruction quality.

Contribution

The authors propose a gradient scaling technique that balances sampling density in NeRFs, removing the necessity of near planes and preventing background artifacts.

Findings

Effective removal of background collapse in NeRFs

Compatible with existing NeRF implementations

No significant computational overhead

Abstract

NeRF acquisition typically requires careful choice of near planes for the different cameras or suffers from background collapse, creating floating artifacts on the edges of the captured scene. The key insight of this work is that background collapse is caused by a higher density of samples in regions near cameras. As a result of this sampling imbalance, near-camera volumes receive significantly more gradients, leading to incorrect density buildup. We propose a gradient scaling approach to counter-balance this sampling imbalance, removing the need for near planes, while preventing background collapse. Our method can be implemented in a few lines, does not induce any significant overhead, and is compatible with most NeRF implementations.