ZeroCost – Practical
Trainer: Guillaume Jacquemet
Overview: ZeroCostDL4Mic is a Google Colab based no-cost toolbox to explore Deep-Learning in Microscopy. The section introduces the use of Deep Learning to analyze microscopy images and latter train the student in using the ZeroCostDL4Mic platform.
Learning outcomes:
After this section you should be able to:
- Apply basic DL-based methods for bioimage analysis.
- Use open source infrastructure/tools for DL-based bioimage analysis
Materials:
- Github repository: https://github.com/HenriquesLab/ZeroCostDL4Mic/wiki
- Data source: https://zenodo.org/records/3713307
- Data used in the practical
- Data required to train StarDist: https://github.com/HenriquesLab/ZeroCostDL4Mic/wiki/Stardist#data-required-to-train-stardist
- https://datasetsearch.research.google.com/
- https://bioimage.io/#/
- ZeroCostDL4Mic paper: https://doi.org/10.1038/s41467-021-22518-0
- ZeroCostDL4Mic – StarDist example training and test dataset: https://zenodo.org/record/3715492#.YoJbUPNByAn
- Commonly used metrics (see Box 1 for segmentation metrics): Avoiding a replication crisis in deep-learning-based bioimage analysis
- Here is another tool for data annotation: https://github.com/BIOP/ijp-LaRoMe
Note that you need a Google account to access the files for the practical