First NL-BioImaging Advanced Image Analysis Course brings Dutch bioimage analysts together

From 21 to 24 September, Erasmus MC in Rotterdam hosted the first NL-BioImaging Advanced Image Analysis Course. Supported by NL-BioImaging and the NVvM, the course brought together 20 participants from across the Netherlands, taught by image analysts from several different institutes.

Article written by Maarten Paul, Imaging Scientist, Leiden UMC, NL-BI

Inspired by the EVOLVE Train-the-Trainer course

The first ideas for the development of this course came from the Data Analysis and Management node of NL-BioImaging, with the goal of elevating the image analysis expertise in The Netherlands. The organization of the course was led by Maarten Paul (LUMC and Leiden University). Input from the EVOLVE Train-the-Trainer course organized by Euro-BioImaging, which Maarten attended, helped to develop this first edition of this multi-day course. Trainers came from institutes across the country, so the course also served to strengthen cross-institute collaboration and build a Dutch bioimage analysis community. 

Chris Lewis (UU) opened the course with a lecture on the Python ecosystem for image analysis. In the afternoon, Martijn Wehrens (UvA) hosted a code-along session on handling and analyzing bioimages in Jupyter notebooks.

On the second day, Pascal de Boer (UMCG) and Joost de Folter (UvA/UMCG) explained the role of data management in image analysis and how OMERO, an image data management system, can be used together with Python for automated image analysis. In the afternoon, Ihor Smal (UU) gave an overview of deep learning approaches for bioimages, explaining the wide range of applications from denoising to foundation models.

The third day started with an interactive session by Johannes Franz (Maastricht University), who showed how napari can be used to quickly annotate image data and how extensions add AI-based assistance. Daniel Franco Barranco (MRC Laboratory of Molecular Biology, Cambridge, UK) joined online to present BiaPy, the Python tool he and his collaborators developed to train AI models on bioimage data easily and reproducibly.

From tools to Next Generation File Formats

On the final day, Joel Lüthi (BioVisionCenter, Zürich, Switzerland) joined online to discuss the latest developments in OME-Zarr and how his team uses it to build reproducible workflows with Fractal. Furthermore, Przemek Krawczyk (Amsterdam UMC) showed the rapidly developing role of agentic AI in bioimage analysis workflows.

During the last two days, participants also worked together on their own data using the tools they had learned about. Their flash talks showed the large variety of bioimaging research in the Netherlands and the different ways it can benefit from advanced image analysis. The participants evaluated the course positively and would recommend the course to colleagues.

All course materials were collected in an online book and are freely available.

The organizers are already planning a second edition next year and hope to grow the Dutch bioimage analysis community even further. Keep an eye on the NVvM newsletter for updates.