Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields

TL;DR

Mip-NeRF 360 extends neural radiance fields to unbounded scenes with a novel scene parameterization and regularizer, significantly improving rendering quality and detail in 360-degree view synthesis.

Contribution

It introduces mip-NeRF 360, a new model that effectively handles unbounded scenes by addressing sampling, aliasing, and scale issues with innovative techniques.

Findings

Reduces mean-squared error by 57% over mip-NeRF

Produces realistic views and detailed depth maps for complex scenes

Effective in 360-degree scene reconstruction

Abstract

Though neural radiance fields (NeRF) have demonstrated impressive view synthesis results on objects and small bounded regions of space, they struggle on "unbounded" scenes, where the camera may point in any direction and content may exist at any distance. In this setting, existing NeRF-like models often produce blurry or low-resolution renderings (due to the unbalanced detail and scale of nearby and distant objects), are slow to train, and may exhibit artifacts due to the inherent ambiguity of the task of reconstructing a large scene from a small set of images. We present an extension of mip-NeRF (a NeRF variant that addresses sampling and aliasing) that uses a non-linear scene parameterization, online distillation, and a novel distortion-based regularizer to overcome the challenges presented by unbounded scenes. Our model, which we dub "mip-NeRF 360" as we target scenes in which the camera rotates 360 degrees around a point, reduces mean-squared error by 57% compared to mip-NeRF, and is able to produce realistic synthesized views and detailed depth maps for highly intricate, unbounded real-world scenes.