Deep Learning Interior Tomography for Region-of-Interest Reconstruction

TL;DR

This paper introduces a deep learning method for interior tomography that effectively removes artifacts and null space signals from FBP reconstructions, achieving high-quality ROI images with reduced computational cost.

Contribution

The authors propose a novel deep learning architecture specifically designed for ROI reconstruction in interior tomography, improving image quality and computational efficiency.

Findings

Achieves 7-10 dB PSNR improvement over existing methods.

Provides near-perfect ROI reconstruction with reduced artifacts.

Operates with significantly lower run-time complexity.

Abstract

Interior tomography for the region-of-interest (ROI) imaging has advantages of using a small detector and reducing X-ray radiation dose. However, standard analytic reconstruction suffers from severe cupping artifacts due to existence of null space in the truncated Radon transform. Existing penalized reconstruction methods may address this problem but they require extensive computations due to the iterative reconstruction. Inspired by the recent deep learning approaches to low-dose and sparse view CT, here we propose a deep learning architecture that removes null space signals from the FBP reconstruction. Experimental results have shown that the proposed method provides near-perfect reconstruction with about 7-10 dB improvement in PSNR over existing methods in spite of significantly reduced run-time complexity.