Conditional Invertible Neural Networks for Medical Imaging

TL;DR

This paper explores the use of invertible neural networks, specifically flow-based models, to improve uncertainty quantification in medical imaging tasks like low-dose CT and MRI, demonstrating architecture variations and distribution choices impact results.

Contribution

It introduces the application of flow-based invertible neural networks to medical imaging inverse problems, emphasizing uncertainty estimation and architecture optimization.

Findings

Radial base distributions can enhance reconstruction quality.

Flow-based models provide uncertainty estimates in medical imaging.

Architecture choices significantly affect model performance.

Abstract

Over the last years, deep learning methods have become an increasingly popular choice to solve tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the analysis of ill-posed inverse problems, the study of uncertainties is essential. In our work, we apply generative flow-based models based on invertible neural networks to two challenging medical imaging tasks, i.e. low-dose computed tomography and accelerated medical resonance imaging. We test different architectures of invertible neural networks and provide extensive ablation studies. In most applications, a standard Gaussian is used as the base distribution for a flow-based model. Our results show that the choice of a radial distribution can improve the quality of reconstructions.