Tomviz 3.0: A Reimagined Open Source Pipeline for Tomographic Data Analysis

Tomographic microscopy analyses are often spread across multiple applications and custom scripts, making results difficult to reproduce, compare, and share. Parameter tuning frequently requires rerunning an entire analysis, while integrating modern AI methods can introduce complex software dependencies that are difficult to manage.

Tomviz 3.0 is the largest release in the project’s history. Built around a redesigned node-based pipeline, it enables researchers to create reproducible analyses with branching, breakpoints, cached intermediate results, and support for external transforms running in dedicated conda environments. The release also introduces native multi-array support, GPU-accelerated reconstruction methods, and is built on the latest ParaView.

During this webinar, Kitware’s experts will:

• Demonstrate a complete tomography workflow, including data import, alignment, GPU-accelerated reconstruction, segmentation, and interactive 3D visualization.
• Show native multi-array support in action, applying transforms across many arrays at once and coloring by any of them, such as one array per element in an X-ray fluorescence map.
• Show how the new node-based pipeline supports branching, breakpoints, and cached intermediate results to simplify parameter tuning, with built-in plotting such as Fourier Shell Correlation and similarity metrics for comparing algorithms quantitatively.
• Demonstrate how AI segmentation transforms, including SAM 2 and SAM 3, run alongside classical methods in dedicated conda environments through the external pipeline architecture; the same mechanism facilities use to add their own data sources.
• Explore reproducible workflows using state files and pipeline templates, and preview what comes next: trame-based web analysis and deeper multivariate analysis.

Key Takeaways

• Build complete tomography pipelines for reconstruction, visualization, and segmentation in Tomviz 3.0.
• Compare reconstruction methods using branching, breakpoints, cached intermediate results, and built-in plotting.
• Work with multi-element and other multi-array datasets natively throughout the pipeline.
• Run AI segmentation transforms in dedicated conda environments alongside traditional tomography methods.
• Save and share reproducible analyses while exploring upcoming web-based and AI-assisted capabilities.

Who Should Attend

This webinar is designed for:

• Electron microscopists, materials scientists, and researchers working with S/TEM tomography.
• Graduate students, postdoctoral researchers, staff scientists, and microscopy facility or core directors.
• Synchrotron beamline scientists and facility users working with XRF, ptychography, and micro-CT tomography.
• Developers, engineers, and researchers using Tomviz, ParaView, ImageJ/Fiji, IMOD/Etomo, TomoPy, ASTRA, HyperSpy, or napari.
• Applications engineers, product managers, principal investigators, and lab managers evaluating tomography analysis software.