Resolution-Invariant Image Classification Based on Fourier Neural Operators

Kabri S, Roith T, Tenbrinck D, Burger M (2023)


Publication Type: Conference contribution

Publication year: 2023

Journal

Original Authors: Samira Kabri, Tim Roith, Daniel Tenbrinck, Martin Burger

Series: International Conference, SSVM

Book Volume: LNCS. volume 14009

Pages Range: 236-249

Conference Proceedings Title: Scale Space and Variational Methods in Computer Vision

Event location: Santa Margherita di Pula IT

ISBN: 9783031319747

DOI: 10.1007/978-3-031-31975-4_18

Abstract

In this paper we investigate the use of Fourier Neural Operators (FNOs) for image classification in comparison to standard Convolutional Neural Networks (CNNs). Neural operators are a discretization-invariant generalization of neural networks to approximate operators between infinite dimensional function spaces. FNOs—which are neural operators with a specific parametrization—have been applied successfully in the context of parametric PDEs. We derive the FNO architecture as an example for continuous and Fréchet-differentiable neural operators on Lebesgue spaces. We further show how CNNs can be converted into FNOs and vice versa and propose an interpolation-equivariant adaptation of the architecture.

Authors with CRIS profile

Involved external institutions

How to cite

APA:

Kabri, S., Roith, T., Tenbrinck, D., & Burger, M. (2023). Resolution-Invariant Image Classification Based on Fourier Neural Operators. In Scale Space and Variational Methods in Computer Vision (pp. 236-249). Santa Margherita di Pula, IT.

MLA:

Kabri, Samira, et al. "Resolution-Invariant Image Classification Based on Fourier Neural Operators." Proceedings of the International Conference on Scale Space and Variational Methods in Computer Vision, Santa Margherita di Pula 2023. 236-249.

BibTeX: Download