Learned SIRT for Cone Beam Computed Tomography Reconstruction

TL;DR

The paper presents a deep learning-based iterative reconstruction method called learned SIRT for cone beam CT, demonstrating superior image quality over traditional methods and scalability to real-world clinical data.

Contribution

It introduces a novel learned SIRT algorithm that combines model knowledge with deep learning for improved 3D cone beam CT reconstruction.

Findings

Achieves higher PSNR and SSIM than FBP, SIRT, and U-net post-processing.

Performs comparably to learned primal dual algorithm in 2D reconstructions.

Scales effectively to clinical problems and physical phantom measurements.

Abstract

We introduce the learned simultaneous iterative reconstruction technique (SIRT) for tomographic reconstruction. The learned SIRT algorithm is a deep learning based reconstruction method combining model knowledge with a learned component. The algorithm is trained by mapping raw measured data to the reconstruction results over several iterations. The Learned SIRT algorithm is applied to a cone beam geometry on a circular orbit, a challenging problem for learned methods due to its 3D geometry and its inherent inability to completely capture the patient anatomy. A comparison of 2D reconstructions is shown, where the learned SIRT approach produces reconstructions with superior peak signal to noise ratio (PSNR) and structural similarity (SSIM), compared to FBP, SIRT and U-net post-processing and similar PSNR and SSIM compared to the learned primal dual algorithm. Similar results are shown for cone beam geometry reconstructions of a 3D Shepp Logan phantom, where we obtain between 9.9 and 28.1 dB improvement over FBP with a substantial improvement in SSIM. Finally we show that our algorithm scales to clinically relevant problems, and performs well when applied to measurements of a physical phantom.