Unsupervised learning approaches to characterizing heterogeneous samples using X-ray single-particle imaging

Zhuang, Y.; Awel, S.; Barty, A.; Bean, R.; Bielecki, J.; Bergemann, M.; Daurer, B.J.; Ekeberg, T.; Estillore, A.D.; Fangohr, H.; Giewekemeyer, K.; Hunter, M.S.; Karnevskiy, M.; Kirian, R.A.; Kirkwood, H.; Kim, Y.; Koliyadu, J.; Lange, H.; Letrun, R.; Lübke, J.; Mall, A.; Michelat, T.; Morgan, A.J.; Roth, N.; Samanta, A.K.; Sato, T.; Shen, Z.; Sikorski, M.; Schulz, F.; Spence, J.C.H.; Vagovic, P.; Wollweber, T.; Worbs, L.; Xavier, P.L.; Yefanov, O.; Maia, F.R.N.C.; Horke, D.A.; Küpper, J.; Loh, N.D.; Mancuso, A.P.; Chapman, H.N.; Ayyer, K.

doi:10.1107/S2052252521012707

Figure 5
(a) VAE-encoded 1D latent number z plotted against CLPCA-2 for the input sample patterns. The colour coding is the gain parameter of each pattern. (b) A histogram of latent number z. (c) A histogram of CLPCA-2. (d) The volume `evolution' along z, with volumes calculated with a density threshold equal to 10⁻⁴ of the total mass. The dashed horizontal line is the size of the support volume in phase retrieval. Vertical grey lines show the locations of the 12 selected z numbers. (e) The top three rows are the VAE-generated intensity volumes from the 12 selected z numbers on a logarithmic scale, the three rows showing slices in (from row 1 to row 3) the 〈100〉, 〈110〉 and 〈111〉 directions, respectively. The bottom three rows show their corresponding density projections in real space.

IUCrJ

Volume 9| Part 2| March 2022| Pages 204-214

ISSN: 2052-2525

https://doi.org/10.1107/S2052252521012707

PHYSICS | FELS

Open

access

Follow IUCrJ

Search IUCr Journals		doi		Advanced search
Author		volume	page