AI-Based Pipelines for Bioimage Analysis
Date:
In May 2025, I delivered a session titled “AI-Based Pipelines for Bioimage Analysis” at the NEUBIAS Bioimage Analysis Course hosted at Institut Pasteur, Paris. This course is part of the prestigious NEUBIAS training series, focused on advancing bioimage analysis across Europe.
My talk introduced participants to deep learning concepts and tools tailored for microscopy workflows, including:
- When and why to apply deep learning in microscopy
- The lifecycle of a deep learning model
- Transformer architectures and Vision Transformers (ViTs)
- Segment Anything Model (SAM) and its integration into Fiji via SAMJ
- Practical use of tools such as deepImageJ, BiaPy, and BioImage.IO
We also conducted hands-on exercises using SAMJ to annotate biological images efficiently, supported by curated datasets and workflows.
📁 Slides & Materials: Zenodo (10.5281/zenodo.15395889)
💻 GitHub Exercises: Pasteur-BioImage-Analysis-Course-2025
Special thanks to Carlos Javier García-López-de-Haro for his support and contributions to the course content.