Scatter Correction for Spectral CT Using a Primary Modulator Mask

The problem of scattered radiation correction in computed tomography (CT) is well known because scatter induces a bias, a loss of contrast and artifacts. Numerous strategies have been proposed in conventional CT (using energy-integrating detectors) but the problem is still open in the field of spect...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on medical imaging 2020-06, Vol.39 (6), p.2267-2276
Main Authors: Pivot, Odran, Fournier, Clarisse, Tabary, Joachim, Letang, Jean Michel, Rit, Simon
Format: Article
Language:English
Subjects:
Computed tomography
Computer Science
Computer simulation
Cost function
Detectors
Energy
Error correction
Error reduction
Estimation
Imaging techniques
Mathematical model
Medical Imaging
Modulation
Photon counting detectors
Photonics
primary modulator mask
Radiation
scatter correction
Scattering
spectral CT
Spline functions
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The problem of scattered radiation correction in computed tomography (CT) is well known because scatter induces a bias, a loss of contrast and artifacts. Numerous strategies have been proposed in conventional CT (using energy-integrating detectors) but the problem is still open in the field of spectral CT, a new imaging technique based on energy-selective photon counting detectors. The aim of the present study is to introduce a scatter correction method adapted to multi-energy imaging and based on the use of a primary modulator mask. The main contributions are a correction matrix, which compensates for the effect of the mask, a scatter model based on B-splines and a cost function based on the mask structures and robust to the object structures. The performances of the method have been evaluated on both simulated and experimental data. The mean relative error was reduced from 20% in the lower energy-bins without correction to 4% with the proposed technique, which is close to the error caused by statistical noise.
ISSN:0278-0062
1558-254X
DOI:10.1109/TMI.2020.2970296