PyNeRF: Pyramidal Neural Radiance Fields

TL;DR

PyNeRF introduces a multi-resolution grid approach for Neural Radiance Fields, significantly enhancing rendering quality and training speed by explicitly modeling scale variations, and is compatible with existing accelerated NeRF methods.

Contribution

The paper proposes a simple, effective modification to grid-based NeRF models by training multiple model heads at different resolutions, improving scale-aware rendering without complex encodings.

Findings

Reduces rendering error by 20-90% across scenes.

Achieves 60x faster training compared to Mip-NeRF.

Maintains minimal performance overhead during rendering.

Abstract

Neural Radiance Fields (NeRFs) can be dramatically accelerated by spatial grid representations. However, they do not explicitly reason about scale and so introduce aliasing artifacts when reconstructing scenes captured at different camera distances. Mip-NeRF and its extensions propose scale-aware renderers that project volumetric frustums rather than point samples but such approaches rely on positional encodings that are not readily compatible with grid methods. We propose a simple modification to grid-based models by training model heads at different spatial grid resolutions. At render time, we simply use coarser grids to render samples that cover larger volumes. Our method can be easily applied to existing accelerated NeRF methods and significantly improves rendering quality (reducing error rates by 20-90% across synthetic and unbounded real-world scenes) while incurring minimal performance overhead (as each model head is quick to evaluate). Compared to Mip-NeRF, we reduce error rates by 20% while training over 60x faster.