Lose The Views: Limited Angle CT Reconstruction via Implicit Sinogram Completion

TL;DR

This paper introduces a neural network-based method for limited angle CT reconstruction that completes sinograms implicitly, enabling high-quality reconstructions from less-than-180-degree scans, with a confidence measure for reliability.

Contribution

It proposes a novel implicit sinogram completion approach using neural networks, improving reconstruction quality in limited angle CT scenarios and providing a confidence measure without ground truth.

Findings

Outperforms baseline methods in limited angle CT reconstruction

Provides a confidence measure correlating with PSNR

Preserves 3D structure effectively in reconstructions

Abstract

Computed Tomography (CT) reconstruction is a fundamental component to a wide variety of applications ranging from security, to healthcare. The classical techniques require measuring projections, called sinograms, from a full 180$^\circ$ view of the object. This is impractical in a limited angle scenario, when the viewing angle is less than 180$^\circ$, which can occur due to different factors including restrictions on scanning time, limited flexibility of scanner rotation, etc. The sinograms obtained as a result, cause existing techniques to produce highly artifact-laden reconstructions. In this paper, we propose to address this problem through implicit sinogram completion, on a challenging real world dataset containing scans of common checked-in luggage. We propose a system, consisting of 1D and 2D convolutional neural networks, that operates on a limited angle sinogram to directly produce the best estimate of a reconstruction. Next, we use the x-ray transform on this reconstruction to obtain a "completed" sinogram, as if it came from a full 180$^\circ$ measurement. We feed this to standard analytical and iterative reconstruction techniques to obtain the final reconstruction. We show with extensive experimentation that this combined strategy outperforms many competitive baselines. We also propose a measure of confidence for the reconstruction that enables a practitioner to gauge the reliability of a prediction made by our network. We show that this measure is a strong indicator of quality as measured by the PSNR, while not requiring ground truth at test time. Finally, using a segmentation experiment, we show that our reconstruction preserves the 3D structure of objects effectively.