“Neural Field Convolutions by Repeated Differentiation” by Nsampi, Djeacoumar, Seidel, Ritschel and Leimkühler
Conference:
Type(s):
Title:
- Neural Field Convolutions by Repeated Differentiation
Session/Category Title: How To Deal With NERF?
Presenter(s)/Author(s):
Abstract:
Neural fields are evolving towards a general-purpose continuous representation for visual computing. Yet, despite their numerous appealing properties, they are hardly amenable to signal processing. As a remedy, we present a method to perform general continuous convolutions with general continuous signals such as neural fields. Observing that piecewise polynomial kernels reduce to a sparse set of Dirac deltas after repeated differentiation, we leverage convolution identities and train a repeated integral field to efficiently execute large-scale convolutions. We demonstrate our approach on a variety of data modalities and spatially-varying kernels.
Overview Page:
Submit a story:
If you would like to submit a story about this presentation, please contact us: historyarchives@siggraph.org
