Summary: Selective Plane Illumination Microscopy (SPIM) allows to image developing organisms in 3D at unprecedented temporal resolution over long periods of time. The resulting massive amounts of raw image data requires extensive processing interactively via dedicated graphical user interface (GUI) applications. The consecutive processing steps can be easily automated and the individual time points can be processed independently, which lends itself to trivial parallelization on a high performance computing (HPC) cluster. Here, we introduce an automated workflow for processing large multiview, multichannel, multiillumination time-lapse SPIM data on a single workstation or in parallel on a HPC cluster. The pipeline relies on snakemake to resolve dependencies among consecutive processing steps and can be easily adapted to any cluster environment for processing SPIM data in a fraction of the time required to collect it. Availability and implementation: The code is distributed free and open source under the MIT license http://opensource.org/licenses/MIT. The source code can be downloaded from github: https://github.com/mpicbg-scicomp/snakemake-workflows. Documentation can be found here: http://fiji.sc/Automated_workflow_for_parallel_Multiview_Reconstruction. Contact: schmied@mpi-cbg.de Supplementary information: Supplementary data are available at Bioinformatics online.
Accurate characterization of radiation pulse profiles is crucial for optimizing beam quality and enhancing experimental outcomes in free electron laser (FEL) research. In this paper, we present a unique approach that employs machine learning techniques for real-time virtual diagnostics of FEL radiation pulses. Our simple artificial intelligence (AI)-based diagnostic tool utilizes longitudinal phase space data obtained from the X-band transverse deflecting structure to reconstruct the temporal profile of FEL pulses in real time. Unlike traditional single-shot methods, this AI-driven solution provides a noninvasive, highly efficient alternative for pulse characterization. By leveraging state-of-the-art machine learning models, our method facilitates precise, single-shot measurements of FEL pulse power, offering significant advantages for FEL science research. This work outlines the conceptual framework, methodology, and validation results of our virtual diagnostic tool, demonstrating its potential to significantly impact FEL research.
Künstliche Intelligenz als Überbegriff steht schon länger im Fokus von Klinikern, Gesundheitsökonomen und medizinischen Wissenschaftlern. Entdeckungen in diesem weiten Feld sind aber nur vereinzelt im klinischen Alltag sichtbar, da insbesondere die diagnostische Entscheidungshoheit dem Menschen zugestanden wird. Doch steht schon heute fest: Der Einsatz von künstlicher Intelligenz wird nach und nach zu einem Wandel in der Medizin führen.