dani-lbnl/aitomo

Integrating acquisition and AI in tomography - Lorentz Center, Leiden

Real-time visualization and analysis

• Event: Integrating acquisition and AI in tomography at Lorentz Center, Leiden, NE

• Presentation: Team's PDF

Aim: Enabling using deep learning at facilities to produce real-time visualization and analysis during experiments Potential outcomes: A document describing current and future strategies for realizing implementations of real-time visualization and analysis, including tools for annotation which are fundamental for enabling deep learning as even unsupervised methods require annotation to provide evidences that they are optimized to the purpose they are designed for. The expectaction is that these plans will facilitate future collaborations and could lead to several novel methods and scientific publications.

Toward Real-time Visualization and Data Analysis

These are links to Jupyter Notebooks used as part of tutorial delivered at LBNL ALS User Meeting:

• 01_getstarted
• 02_stackoverview
• 03_voxeldistribution
• 04_binarypartition
• 05_multiscale
• 06_tridimvis

Naturalis triceratops

• Download STL to print your triceratops or visualize using itkwidgets

Other resources

• 3D Data Specimens:
  - Fibers - Beads
• Explore HPC-aware viz code
• Python Cluster Configuration: - Coiled.io Cloud - NERSC

Refs

@article{Battery:NPJ:2023,
author = {Huang, Ying and Perlmutter, David and Su, Andrea and Quenum, Jerome and Zenyuk, Iryna and Ushizima, Daniela},
isbn = {2057-3960},
journal = {Nature Computational Materials},
title = {Detecting Lithium Plating Dynamics in a Solid-State Battery with Operando X-ray Computed Tomography using Machine Learning (accepted)},
year = {2023},
}      

@article{Pinto:2022,
author = {Allan Pinto, Gabriel Borin, Bruno Carlos, Matheus L. Bernardi, Matheus F. Sarmento, Alan Z. Peixinho, Thiago V. Spina and Eduardo X. Miqueles},
title = {Annotat3D: A Modern Web Application for Interactive Segmentation of Volumetric Images at Sirius/LNLS},
journal = {Synchrotron Radiation News},
volume = {35},
number = {4},
pages = {36-43},
year = {2022},
publisher = {Taylor & Francis},
doi = {10.1080/08940886.2022.2112501},
}
}      

@InProceedings{SC:2020,
author = {Daniela Ushizima and Matthew McCormick and Dilworth Parkinson},
title = {Accelerating Microstructural Analytics with Dask for Volumetric X-ray Images},
booktitle = {2020 IEEE/ACM Wrkshp on Python for High-Performance and Scientific Computing (PyHPC) at Super Computing},
month = {Nov},
year = {2020},
pages = {41-48},
}