Scatter Correction in X-ray CT by Physics-Inspired Deep Learning

TL;DR

This paper introduces two physics-inspired deep learning methods, PhILSCAT and OV-PhILSCAT, for correcting scatter artifacts in X-ray CT, leveraging both initial reconstructions and measurements to improve image quality.

Contribution

The work presents novel deep learning approaches that incorporate physics-based information and reconstruction data for scatter correction in CT, outperforming previous purely data-driven methods.

Findings

Consistent improvement over existing deep neural network scatter correction methods.

Effective incorporation of initial reconstructions enhances correction accuracy.

Numerical experiments validate the methods using Monte-Carlo simulated data.

Abstract

Scatter due to interaction of photons with the imaged object is a fundamental problem in X-ray Computed Tomography (CT). It manifests as various artifacts in the reconstruction, making its abatement or correction critical for image quality. Despite success in specific settings, hardware-based methods require modification in the hardware, or increase in the scan time or dose. This accounts for the great interest in software-based methods, including Monte-Carlo based scatter estimation, analytical-numerical, and kernel-based methods, with data-driven learning-based approaches demonstrated recently. In this work, two novel physics-inspired deep-learning-based methods, PhILSCAT and OV-PhILSCAT, are proposed. The methods estimate and correct for the scatter in the acquired projection measurements. Different from previous works, they incorporate both an initial reconstruction of the object of interest and the scatter-corrupted measurements related to it, and use a deep neural network architecture and cost function, both specifically tailored to the problem. Numerical experiments with data generated by Monte-Carlo simulations of the imaging of phantoms reveal consistent improvement over a recent purely projection-domain deep neural network scatter correction method.