Improving Generalizability in Limited-Angle CT Reconstruction with Sinogram Extrapolation

TL;DR

This paper introduces ExtraPolationNetwork, a deep learning model with a sinogram extrapolation module that enhances limited-angle CT reconstruction and significantly improves cross-dataset generalization.

Contribution

The paper proposes a theoretically justified sinogram extrapolation module that boosts generalizability of deep learning models in limited-angle CT reconstruction.

Findings

Achieves state-of-the-art performance on NIH-AAPM dataset.

Significantly improves generalization on unseen datasets like COVID-19 and LIDC.

Demonstrates the effectiveness of sinogram extrapolation in limited-angle CT reconstruction.

Abstract

Computed tomography (CT) reconstruction from X-ray projections acquired within a limited angle range is challenging, especially when the angle range is extremely small. Both analytical and iterative models need more projections for effective modeling. Deep learning methods have gained prevalence due to their excellent reconstruction performances, but such success is mainly limited within the same dataset and does not generalize across datasets with different distributions. Hereby we propose ExtraPolationNetwork for limited-angle CT reconstruction via the introduction of a sinogram extrapolation module, which is theoretically justified. The module complements extra sinogram information and boots model generalizability. Extensive experimental results show that our reconstruction model achieves state-of-the-art performance on NIH-AAPM dataset, similar to existing approaches. More importantly, we show that using such a sinogram extrapolation module significantly improves the generalization capability of the model on unseen datasets (e.g., COVID-19 and LIDC datasets) when compared to existing approaches.