AI4Life

AI4Life_Logo
Menu

Month: June 2025

Categories
News

Release of the New BioImage Model Zoo Website

Release of the New BioImage Model Zoo Website

The BioImage Model Zoo unveils a complete redesign—offering a more collaborative and interactive platform for the community, with rapid model testing feedback to enhance the contributor experience.

Built as part of the AI4Life project, the revamped platform emerged from the community feedback focused on serving the research community. Early user input revealed that contributor experience is key to enabling greater model ingestion and community engagement. We therefore rebuilt the website from the ground up, focusing on streamlined uploading, testing, validation, and review workflows powered by scalable infrastructure.

What’s new?

The new website comes with a bunch of new features:

  • Complete frontend rewrite with enhanced user experience and more powerful model search capabilities.
  • Intuitive model uploader with in-browser file editing support, streamlining the contribution process.
  • Instant model testing during the upload process, providing immediate feedback and enhanced quality control for contributors.
  • Collaborative review interface allowing reviewers to quickly edit, review, and approve models with real-time collaboration tools.
  • New Hypha-powered backend featuring advanced artifact management, user login, permission systems, real-time collaboration, and comprehensive APIs for programmatic access—enabling seamless model access and upload.
  • BioEngine server infrastructure powered by Ray, enabling robust model test runs and scalable compute resources. It also allows on-premise deployment with an intuitive wizard for easy setup.
  • Resilient storage infrastructure backed by S3 storage hosted at EMBL-EBI with automatic Zenodo backup for long-term preservation.

These upgrades transform the entire model sharing experience, whether you’re contributing cutting-edge models, discovering tools for your research, or collaborating with the community to advance reproducible AI in bioimaging. Looking ahead, this new foundation enables us to make AI models even more accessible through intuitive web applications and AI Agents, designed for federated deployment across institutions to bring powerful bioimaging AI directly to researchers worldwide.

Explore the new platform: https://bioimage.io

Behind the scenes: The rebuilding process was supercharged by the emergence of AI coding assistants—tools enabled rapid frontend development and backend integration, showcasing how “vibe coding” can accelerate scientific software development.

Categories
News

The AI4Life Challenge #2 is now live

The AI4Life Challenge #2 is now live

We’re happy to announce the second AI4Life Challenge, this time focused on supervised denoising of microscopy images, a key task in enhancing image quality for biomedical research.

Hosted on the Grand Challenge platform, we invite researchers, developers, and enthusiasts to develop solutions for image restoration using deep learning.

Participants will:

  • Access paired datasets of noisy and clean images.
  • Develop and submit models for data-driven, content-aware denoising.
  • Be evaluated against ground truth for quantitative performance.
  • Contribute their solutions to a core task in bioimage analysis.

Whether you’re a seasoned researcher or new to the field, this challenge is an opportunity to test and improve your models on real-world microscopy data, and support FAIR, open science while doing it.

Get started now!

👉 https://ai4life-mdc25.grand-challenge.org/

 

Categories
News

The AI4Life Model Evaluation Platform: building bridges between the BioImage Archive and the BioImage Model Zoo

The AI4Life Model Evaluation Platform: building bridges between the BioImage Archive and the BioImage Model Zoo

by Teresa Zulueta-Coarasa

The growing number of  models available at the BioImage Model Zoo, are a valuable resource for life scientists aiming to analyze complex imaging data using AI. However, users can struggle to identify which model best matches their own data. At the same time, model developers frequently lack access to diverse datasets, limiting their ability to evaluate how well their models generalize across different biological imaging scenarios.

To address both challenges, we have developed the Model Evaluation Platform which provides side-by-side performance comparisons of different models on a wide variety of imaging datasets from the BioImage Archive.

 

The platform offers two complementary ways to explore model performance. In the model-centric view (“Model Evaluation Platform”), you can compare how a single model performs across a wide range of datasets. This is especially useful for model developers, who can assess how their contributions to the BioImage Model Zoo generalise across diverse imaging data. Alternatively, the dataset-centric view (“Analysed Datasets”) allows you to examine how different models perform on the same dataset. This perspective is ideal for life scientists, who can identify datasets similar to their own and discover which models are likely to yield the most relevant results for their research. This dual approach makes it easier for both scientists and developers to identify optimal use cases for each model.

When you navigate to either the “Analysed Datasets” or the “Model Evaluation Platform”, you will find the models or datasets available within that collection, helping you quickly understand what’s included. From there, you can drill down into individual model or dataset views, which offer more detailed summaries.

On datasets with available ground truth annotations, you will be able to find metrics—such as precision, recall, Intersection over Union (IoU), and Dice score— that you can use to compare the performance of models.

These pages also feature graphical visualisations of model inputs, outputs, and ground truth references—the data used to compute the numerical scores. Input and output images are also available for datasets that do not have ground truth annotations, allowing you to visually inspect the performance of the models.

This Model Evaluation Platform is the first step to achieve our long-term goal to continually increase the utility of the Model Zoo by supporting end-users in selecting appropriate models and enabling developers to benchmark their models against current state-of-the-art approaches with ease.



Categories
News

AI4Life Community Event Celebrating AI Innovation in Life Sciences held in Helsinki

AI4Life Community Event Celebrating AI Innovation in Life Sciences held in Helsinki

The AI4Life project organised a community event on May 27–28, 2025, at Katajanokan Kasino in Helsinki. 

AI4Life brings together researchers, developers, infrastructure providers, and industry stakeholders from across Europe to accelerate the use of AI in biological and biomedical imaging. Over the course of two days, the event showcased cutting-edge AI tools, impactful research collaborations, and the vibrant community that AI4Life has fostered. 

Highlights included interactive workshops for both life scientists and AI developers, offering hands-on sessions with the BioImage Model Zoo (BMZ)—a platform hosting ready-to-use pre-trained AI models for microscopy image analysis. Attendees explored how to make image data and annotations standardised and FAIR (Findable, Accessible, Interoperable, and Reusable). The BioImage Archive hosts AI-ready datasets that have been benchmarked using AI models from the BMZ. On the infrastructure side, BioEngine was shown to deploy AI models on-premise.

The event also brought together experts from other communities that AI4Life collaborates with, such as the Global BioImage Analysts’ Society  (GloBIAS), in which several AI4Life members are volunteers, and AI4EOSC, a European project that has worked in the interoperability of AI models and that is hosting the AI4Life Cloud Marketplace.

Participants also learned about real-world applications from the AI4Life open calls use cases, such as 3D electron microscopy for plankton research and resolution-enhancement techniques in intravital microscopy. The Finnish Advanced Microscopy Node (FiAM) also promoted its image analysis capabilities. Success stories from the industry partner Zeiss and community contributors, like BiaPy, QuPath, and SpotMAX, demonstrated the project’s impact on both academia and industry.

The event concluded with a forward-looking discussion on the strategic role of research infrastructures, with presentations from EMPHASIS ERIC and EU-OPENSCREEN ERIC on the evolving landscape of European Open Science and AI policy. The discussion emphasised AI4Life’s contributions to interoperability and responsible data sharing.

Funded by the European Union’s Horizon Europe research and innovation programme (grant agreement no. 101057970), AI4Life has laid essential foundations for the future of open, AI-driven science. The tools, datasets, and knowledge produced by AI4Life will continue to benefit the global research community via platforms like the BioImage Model Zoo and the BioImage Archive.