Open AccessPost‐reconstruction 3D single‐distance phase retrieval for multi‐stage phase‐contrast tomography with photon‐counting detectors
Author(s) -
Brun Francesco,
Brombal Luca,
Di Trapani Vittorio,
Delogu Pasquale,
Donato Sandro,
Dreossi Diego,
Rigon Luigi,
Longo Renata
Publication year - 2019
Publication title -
journal of synchrotron radiation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.172
H-Index - 99
ISSN - 1600-5775
DOI - 10.1107/s1600577519000237
Subject(s) - optics , context (archaeology) , detector , phase retrieval , iterative reconstruction , projection (relational algebra) , computer science , pipeline (software) , image stitching , tomography , phase contrast imaging , computer vision , artificial intelligence , contrast (vision) , image resolution , image quality , physics , algorithm , phase contrast microscopy , image (mathematics) , geology , paleontology , fourier transform , quantum mechanics , programming language
In the case of single‐distance propagation‐based phase‐contrast X‐ray computed tomography with synchrotron radiation, the conventional reconstruction pipeline includes an independent 2D phase retrieval filtering of each acquired projection prior to the actual reconstruction. In order to compensate for the limited height of the X‐ray beam or the small sensitive area of most modern X‐ray photon‐counting detectors, it is quite common to image large objects with a multi‐stage approach, i.e. several acquisitions at different vertical positions of the sample. In this context, the conventional reconstruction pipeline may introduce artifacts at the margins of each vertical stage. This article presents a modified computational protocol where a post‐reconstruction 3D volume phase retrieval is applied. By comparing the conventional 2D and the proposed 3D reconstructions of a large mastectomy specimen (9 cm in diameter and 3 cm in height), it is here shown that the 3D approach compensates for the multi‐stage artifacts, it avoids refined projection stitching, and the image quality in terms of spatial resolution, contrast and contrast‐to‐noise ratio is preserved.