Accelerated Optimization of Implicit Neural Representations for CT Reconstruction

Document Type

Conference Proceeding

Publication Date

2025

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Source Publication

2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI)

Source ISSN

979-8-3315-2053-3

Original Item ID

DOI: 10.1109/ISBI60581.2025.10981244

Abstract

Inspired by their success in solving challenging inverse prob-lems in computer vision, implicit neural representations (INRs) have been recently proposed for reconstruction in low-dose/sparse-view X-ray computed tomography (CT). An INR represents a CT image as a small-scale neural network that takes spatial coordinates as inputs and outputs attenuation values. Fitting an INR to sinogram data is similar to classical model-based iterative reconstruction methods. However, training INRs with losses and gradient-based algorithms can be prohibitively slow, taking many thousands of iterations to converge. This paper investigates strategies to accelerate the optimization of INRs for CT reconstruction. In particular, we propose two approaches: (1) using a modified loss function with improved conditioning, and (2) an algorithm based on the alternating direction method of multipliers. We illustrate that both of these approaches significantly accelerate INR-based reconstruction of a synthetic breast CT phantom in a sparse-view setting.

Comments

Published as part of the proceedings of the 2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI). DOI.

Link to Full Text

Share

COinS