Neural Vector Tomography for Reconstructing a Magnetization Vector Field

TL;DR

This paper introduces a neural field-based approach for vector tomography that produces high-quality, noise-robust reconstructions of magnetization vector fields, outperforming traditional discretized methods especially with symmetry considerations.

Contribution

The authors propose a novel neural vector field modeling technique for tomography that enhances reconstruction quality and noise robustness compared to existing discretized methods.

Findings

Neural vector fields improve reconstruction quality under noise.

The method outperforms traditional techniques in the presence of symmetry.

High-quality reconstructions are achieved without pixelation artifacts.

Abstract

Discretized techniques for vector tomographic reconstructions are prone to producing artifacts in the reconstructions. The quality of these reconstructions may further deteriorate as the amount of noise increases. In this work, we instead model the underlying vector fields using smooth neural fields. Owing to the fact that the activation functions in the neural network may be chosen to be smooth and the domain is no longer pixelated, the model results in high-quality reconstructions, even under presence of noise. In the case where we have underlying global continuous symmetry, we find that the neural network substantially improves the accuracy of the reconstruction over the existing techniques.