view article
Figure 4
Input images (a) and separation results (b–g) of the ZP/NiO dataset. Like in the cases shown in Fig. 3[link], the results were obtained with f1/2 set to either single filters or shallow DIPs. The final values of α1|α2 are indicated at the upper right corners of the corresponding subplots. While the influence of γexcl on slice 1 is minimal, it greatly affects the balance between separation effectiveness and image resolution for slice 2. The rightmost column shows the normalized power spectra of slice 2, plotted as (log10P)2.5 (where P is the original power spectrum). A slanted streak corresponding to the periodicity of the zone plate's ghost features can be seen obviously in the input image's spectrum. In the outputs of the DDIP (i.e. y2), the spectrum density of this streak becomes much lower. Also, we again see that, when we use shallow DIPs for f1/2 and over-weigh the exclusion loss, a smaller α1/α2 ratio is yielded.

Journal logoJOURNAL OF
SYNCHROTRON
RADIATION
ISSN: 1600-5775
Volume 28| Part 4| July 2021| Pages 1137-1145
https://doi.org/10.1107/S1600577521003507
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