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.