Spectral Prefiltering of Neural Fields

SIGGRAPH Asia 2025 (Conference Papers)
Mustafa B. Yaldiz, Ishit Mehta, Nithin Raghavan, Andreas Meuleman*, Tzu-Mao Li, Ravi Ramamoorthi
University of California San Diego • * Inria & Université Côte d’Azur (France)
📄 Paper (PDF) 💻 Code ▶️ Video
Teaser
We present a training method for neural fields that enables linear prefiltering with multiple reconstruction filters. At training time, the neural field sees parameters of a single symmetric filter. At test time, we support prefiltering a variety of unseen filters (e.g., Box or Lanczos). Here, we show neural fields trained on an image (with bottom-right insets of frequency spectrum) and signed distance function using Gaussian filters, with generalization on Box and Lanczos filters. Images from Adobe FiveK; © original photographers/Adobe. Mesh models from the Stanford 3D Scanning Repository; © Stanford Computer Graphics Laboratory.

Abstract

Neural fields excel at representing continuous visual signals but typically operate at a single, fixed resolution. We present a simple yet powerful method to optimize neural fields that can be prefiltered in a single forward pass. Key innovations and features include: (1) We perform convolutional filtering in the input domain by analytically scaling Fourier feature embeddings with the filter’s frequency response. (2) This closed-form modulation generalizes beyond Gaussian filtering and supports other parametric filters (Box and Lanczos) that are unseen at training time. (3) We train the neural field using single-sample Monte Carlo estimates of the filtered signal. Our method is fast during both training and inference, and imposes no additional constraints on the network architecture. We show quantitative and qualitative improvements over existing methods for neural-field filtering.

Comparisons against Neural Gaussian Scale-Space Fields

Expand each scene for interactive comparison

Gaussian filter

Alien

Interactive zoom

PM5644

Interactive zoom

CuriousCat

Interactive zoom

ParkCafe

Interactive zoom

Box filter

Alien

Interactive zoom

PM5644

Interactive zoom

CuriousCat

Interactive zoom

Parkcafe

Interactive zoom

Lanczos

Alien

Interactive zoom

PM5644

Interactive zoom

CuriousCat

Interactive zoom

ParkCafe

Interactive zoom

BibTeX

@inproceedings{Coming soon...}

Acknowledgements

The website style is inspired by Michaël Gharbi and Ref‑NeRF. We thank collaborators and colleagues for discussions and feedback.