view article
Figure 3
Comparison of our masked phase retrieval method with alternative approaches, shown using sagittal cross-sections of the rabbit head CT reconstruction volumes from Section 3[link]. Panels (a) and (d) are two cross-sections of the phase contrast CT as a baseline for the other algorithms. Panels (b) and (e) are the bone–tissue interface phase retrieved image slices, equation (3)[link], showing considerable remnant noise. Panels (c) and (f) show the result of applying equation (1)[link] phase retrieval to each projection under the assumption that only brain tissue is present, leading to over-blurring completely obscuring brain tissue features in proximity to the bone. Panels (h) and (i) are retrieved using 3DMPR to achieve an equally low SNR image whilst correctly reconstructing the A/B material interface. Brown boxes denote regions used for SNR analysis while the white arrow in (h) highlights a gap in the animal skull that is not visible in the equation (1)[link] phase retrieval. All images use the same colour palette with the region surrounding the animal filled with agarose.

Journal logoJOURNAL OF
SYNCHROTRON
RADIATION
ISSN: 1600-5775
Volume 32| Part 5| September 2025| Pages 1319-1327
https://doi.org/10.1107/S1600577525005843
Follow J. Synchrotron Rad.
Sign up for e-alerts
E-alerts
Follow J. Synchrotron Rad. on Twitter
Twitter
Follow us on facebook
Facebook
Sign up for RSS feeds
RSS