view article
Figure 1
Water-filled versus air-filled xylem vessels revealed through AI-driven segmentation. By leveraging machine learning as part of the Genesis Mission and large-scale HPC inference, segmentation time is reduced from approximately one month of manual annotation to ∼8 minutes, enabling rapid, in situ analysis of plant hydraulic function during experiments. Plant sample provided by Andrew McElrone (UC Davies), data collected at beamline 8.3.2 at the Advanced Light Source, Berkeley Lab.

IUCrJ
Volume 13| Part 3| May 2026| Pages 211-214
ISSN: 2052-2525
https://doi.org/10.1107/S2052252526004161