Lune

CVPR2024Top-tier venue

Alpha Invariance: On Inverse Scaling Between Distance and Volume Density in Neural Radiance Fields

Joshua Ahn, Haochen Wang, Raymond A. Yeh, Greg Shakhnarovich

Abstract

Scale-ambiguity in 3D scene dimensions leads to magnitude-ambiguity of volumetric densities in neural radiance fields, i.e., the densities double when scene size is halved, and vice versa. We call this property alpha invariance. For NeRFs to better maintain alpha invariance, we recommend 1) parameterizing both distance and volume densities in log space, and 2) a discretization-agnostic initialization strategy to guarantee high ray transmittance. We revisit a few popular radiance field models and find that these systems use various heuristics to deal with issues arising from scene scaling. We test their behaviors and show our recipe to be more robust. Visit our project page at https://pals.ttic.edu/p/alpha-invariance . Figure 1. A discretized view of volume rendering. Top: a ray is cut into intervals, each with a density i 0 and interval length di. Bottom: illustration of the weight given to the 3rd interval, computed through alpha compositing. The rendered color is obtained by weighting all the interval colors with their wis. If we scale each di by a constant k, scaling i by 1 k renders the identical color.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Instant neural graphics primitives with a multiresolution hash encoding cites Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
  • 3D Gaussian Splatting for Real-Time Radiance Field Rendering cites Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
  • 3D Gaussian Splatting for Real-Time Radiance Field Rendering cites Instant neural graphics primitives with a multiresolution hash encoding
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on15

  1. 3D Gaussian Splatting for Real-Time Radiance Field RenderingBernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, George DrettakisSIGGRAPH 2023 · 5,687 citations
  2. Instant neural graphics primitives with a multiresolution hash encodingThomas Müller, Alex Evans, Christoph Schied, Alexander KellerSIGGRAPH 2022 · 4,089 citations
  3. Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance FieldsJonathan T. Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman et al.ICCV 2021 · 2,700 citations
  4. NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view ReconstructionPeng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt et al.NeurIPS 2021 · 2,500 citations
  5. Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance FieldsJonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan et al.CVPR 2022 · 1,603 citations

Related papers

  1. VDN-NeRF: Resolving Shape-Radiance Ambiguity via View-Dependence NormalizationBingfan Zhu, Yanchao Yang, Xulong Wang, Youyi Zheng et al.CVPR 2023
  2. NeRF Revisited: Fixing Quadrature Instability in Volume RenderingMikaela Angelina Uy, Kiyohiro Nakayama, Guandao Yang, Rahul Krishna Thomas et al.NeurIPS 2023 · 14 citations
  3. PyNeRF: Pyramidal Neural Radiance FieldsHaithem Turki, Michael Zollhöfer, Christian Richardt, Deva RamananNeurIPS 2023 · 31 citations
  4. Reducing Shape-Radiance Ambiguity in Radiance Fields with a Closed-Form Color Estimation MethodQihang Fang, Yafei Song, Keqiang Li, Liefeng BoNeurIPS 2023 · 8 citations
  5. ABLE-NeRF: Attention-Based Rendering with Learnable Embeddings for Neural Radiance FieldZhe Jun Tang, Tat-Jen Cham, Haiyu ZhaoCVPR 2023
Dusk over the sea between two cliffs drawn in fine vertical lines