A Soft-Threshold Filtering Approach for Tomography Reconstruction from a Limited Number of Projections with Bilateral Edge Preservation.

Tiago Wirtti,Evandro Salles

doi:10.3390/s19102346

Abstract

In X-ray tomography image reconstruction, one of the most successful approaches involves a statistical approach with norm for fidelity function and some regularization function with norm, . Among them stands out, both for its results and the computational performance, a technique that involves the alternating minimization of an objective function with norm for fidelity and a regularization term that uses discrete gradient transform (DGT) sparse transformation minimized by total variation (TV). This work proposes an improvement to the reconstruction process by adding a bilateral edge-preserving (BEP) regularization term to the objective function. BEP is a noise reduction method and has the purpose of adaptively eliminating noise in the initial phase of reconstruction. The addition of BEP improves optimization of the fidelity term and, as a consequence, improves the result of DGT minimization by total variation. For reconstructions with a limited number of projections (low-dose reconstruction), the proposed method can achieve higher peak signal-to-noise ratio (PSNR) and structural similarity index measurement (SSIM) results because it can better control the noise in the initial processing phase.

Highlights

X-ray computer tomography (CT) measures the attenuation of X-ray beams passing through an object, generating projections
In favor of a better understanding of the purpose of this work, we first present a base solution that uses simultaneous algebraic reconstruction technique (SART) reconstruction regularized via total variation (TV) minimization of the discrete gradient transform (DGT) function (SART+DGT), highlighting the relevant parts, and we present our approach
It is worth remembering that x is the image we intend to reconstruct from the noise signal y and, in the modeling process presented in Sections 2 and 3, we use the variable μ to represent it

Summary

Introduction

X-ray computer tomography (CT) measures the attenuation of X-ray beams passing through an object, generating projections. Such projections are processed, resulting in an image (slice) of the examined object. This is known as a CT image reconstruction. The ALARA principle states that only the minimum amount of radiation must be applied to the patient. For this reason, ALARA is widely accepted in the medical CT community [1]. To reduce the X-ray dose of the patient during the CT scan, there are two possibilities: (1) reduce the amount of projection

Methods

Results

Conclusion