Computed Tomography Reconstruction using Generative Energy-Based Priors

TL;DR

This paper introduces a novel unsupervised energy-based regularizer for CT reconstruction that captures domain statistics, improves image quality in limited data scenarios, and offers a probabilistic framework for uncertainty quantification.

Contribution

It presents a new parametric regularizer learned via likelihood maximization, applicable across CT problems, and providing interpretability and uncertainty estimation.

Findings

Outperforms traditional algorithms in limited-angle CT reconstruction

Successfully synthesizes realistic CT images demonstrating learned domain statistics

Provides a flexible, probabilistic framework for improved reconstruction quality

Abstract

In the past decades, Computed Tomography (CT) has established itself as one of the most important imaging techniques in medicine. Today, the applicability of CT is only limited by the deposited radiation dose, reduction of which manifests in noisy or incomplete measurements. Thus, the need for robust reconstruction algorithms arises. In this work, we learn a parametric regularizer with a global receptive field by maximizing it's likelihood on reference CT data. Due to this unsupervised learning strategy, our trained regularizer truly represents higher-level domain statistics, which we empirically demonstrate by synthesizing CT images. Moreover, this regularizer can easily be applied to different CT reconstruction problems by embedding it in a variational framework, which increases flexibility and interpretability compared to feed-forward learning-based approaches. In addition, the accompanying probabilistic perspective enables experts to explore the full posterior distribution and may quantify uncertainty of the reconstruction approach. We apply the regularizer to limited-angle and few-view CT reconstruction problems, where it outperforms traditional reconstruction algorithms by a large margin.