Speaker
Description
At ESRF, tomography data processing is supported by a modular software ecosystem designed to connect acquisition, raw data standardization, reconstruction, visualization, automation, and online data availability. Tomotools, brings together several complementary components used across ESRF tomography beamlines to provide users with reliable and reproducible processing workflows.
The chain starts at acquisition level with Bliss-tomo, which integrates tomography procedures into the ESRF Bliss beamline control system. Acquisition metadata, scan descriptions, motor positions, detector information, and experimental context are then used to provide all the information required to perform any downstream processing. Tools such as nxtomo layer provide a common NeXus-based representation of raw tomography data, making the datasets easier to process and exchange among facilities.
Reconstruction computations is handled by nabu, ESRF’s high-performance tomography reconstruction software, which provides key processing features such as dark/flat-field correction, normalization, center-of-rotation estimation, Paganin phase retrieval, helical reconstruction, stitching, and GPU-accelerated volume reconstruction. Around these core processing tools, tomwer provides graphical interfaces and user-oriented workflows, while ewokstomo enables automated and reproducible execution of tomography processing pipelines. Installed on six ESRF beamlines, ewokstomo connects Bliss acquisition metadata with processing workflows and ESRF data services.
Workflow definitions, parameters, software versions, and input/output data references are saved to provide full data provenance. The final goal is to make reconstructed volumes available to users through the ESRF Data Portal together with the corresponding raw data, acquisition metadata, and processing context. In this way, tomography processing at ESRF becomes traceable, scalable, and increasingly integrated into the beamline data lifecycle.
