Premium
TH‐A‐213CD‐04: A Bone Artifact Reduction Algorithm for Differential Phase‐Contrast CT Based On Statistical Iterative Reconstruction
Author(s) -
Hahn D,
Thibault P,
Noel P,
Bech M,
Pfeiffer F
Publication year - 2012
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4736266
Subject(s) - iterative reconstruction , streaking , pixel , computer science , tomography , reduction (mathematics) , artificial intelligence , tomographic reconstruction , contrast (vision) , medical imaging , projection (relational algebra) , optics , materials science , computer vision , algorithm , physics , mathematics , geometry
Purpose: The purpose of this work is a reduction of the influence of dense materials, e.g. bones, on the reconstruction of grating‐based differential phase‐contrast computed tomography (PCCT). These strongly absorbing materials introduce streaking artifacts, in look and feel similar to metal artifacts in conventional absorption CT. Methods: An iterative reconstruction algorithm, which statistically models the differential phase‐ contrast imaging process, was developed. It allows for individually introducing a statistical weight for each sinogram pixel, i.e. its influence on the resulting reconstruction. In grating‐based imaging the absorption and phase‐contrast signals are simultaneously retrieved from the same set of raw projection images and are thus inherently registered. This can be utilized for getting information on the position of highly dense materials, e.g. bones, from the absorption signal. The statistical weights are then modified according to this information to reduce the influence of sinogram pixels that contain dense material and are likely to cause artifacts. This concept was applied to a tomographic dataset of a mouse acquired with a grating interferometer at the European Synchrotron Radiation Facility in Grenoble, France. Reconstructed with a standard filtered backprojection, the sample dataset suffers from very strong streaking artifacts in the region surrounding the spine. Detail contrast in the soft tissue around the spine is lost due to these streaks. Results: Comparing reconstructions performed with a standard filtered backprojection and with the newly developed algorithm shows a significant reduction of streaking artifacts and a strong improvement in soft tissue contrast in regions influenced by dense material. Conclusions: This work significantly expands the potential of grating‐based differential phase‐contrast computed tomography. It is now possible to offer the high soft‐tissue contrast of phase‐contrast imaging in cases where dense materials such as bones, are present. D.H., P.T., M.B. and F.P. acknowledge financial support through the DFG Cluster of Excellence Munich‐Centre for Advanced Photonics and the European Research Council (FP7, Starting grant #240142).