AI4Life

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.

Categories
News

The BioImage Model Zoo meets the Galaxy platform

The BioImage Model Zoo meets the Galaxy platform

A new Galaxy integration enables researchers to access AI models from the BioImage Model Zoo (BioImage.IO). This integration bridges open-source AI models and an accessible, workflow-based computational environment, allowing scientists across disciplines to include deep learning models in their workflows.

This integration is a leap forward for the AI4Life project, which aims to democratize AI in life sciences. “We’re broadening our focus to engage a wider community of users by expanding the reach of the BioImage.IO Model Zoo,” explains Diana Chiang Jurado, who, together with Leonid Kostrykin, developed a dedicated tutorial to support users. “Using the Galaxy infrastructure, we’re not only lowering technical barriers but also making sure that researchers with no access to local computational resources can still run their analyses.”

Galaxy’s established ecosystem brings some benefits: cloud-based computing, reproducibility through workflow provenance, and a collaborative environment thousands of scientists use worldwide.

Integrating these models wasn’t without its challenges. “The BioImage Model Zoo hosts models built on different architectures and frameworks,” notes Anup Kumar, the leading developer behind the integration. “One of the most exciting outcomes,” Anup adds, “is the accessibility it provides. Scientists who don’t have a background in AI, or who work at underfunded institutions, can now use powerful models with just a few clicks. That’s the kind of impact we’re aiming for.”

With AI-driven image analysis becoming increasingly relevant across scientific fields, from biology to earth sciences or astronomy, the integration sets the stage for new cross-disciplinary collaborations like those outlined in the OSCARS-FIESTA project. “Workflows created in Galaxy are easy to share and adapt,” says Leonid. “That means a model trained on biological data could inspire solutions in climate science or vice versa. It’s about breaking down silos.”

By combining rich model repositories with user-friendly, FAIR workflows, this integration makes advanced image analysis more approachable, adaptable, and impactful for scientists everywhere.

Try the tutorial developed by Diana and Leonid here: Using BioImage.IO models for image analysis in Galaxy.

Categories
News

AI4Life Workshop at the Euro-BioImaging All Hands Meeting 2025

AI4Life Workshop at the Euro-BioImaging All Hands Meeting 2025

On March 25, 2025, AI4Life hosted a dedicated workshop at the Euro-BioImaging All Hands meeting to strengthen collaboration between AI4Life and the Euro-BioImaging Nodes. The session brought together imaging experts, bioimaging facility representatives, and AI researchers to explore how AI-powered tools can enhance image analysis workflows and identify opportunities for joint initiatives.

Objectives

The workshop aimed to:

  • Introduce AI4Life to the Euro-BioImaging Nodes, showcasing its goals, resources, and services.
  • Gather feedback from the Nodes on how AI4Life can further improve its tools and services to better support the bioimaging community.
  • Define actionable collaboration pathways between AI4Life and the Euro-BioImaging Nodes.
  • Establish an ongoing framework for interaction, contribution, and resource exchange.

Discussion

Challenges in Running AI Models Across Different Infrastructures

Participants discussed the difficulties of deploying AI models across various computational infrastructures, such as high-performance computing (HPC) clusters. The importance of standardized environments was highlighted, with Docker containers and tools like DL4MicEverywhere identified as potential solutions to ensure model compatibility across platforms.

Quality Control and Model Validation

Ensuring AI model quality and reliability was a major discussion point. Participants emphasized the need for robust validation processes and user feedback mechanisms to assess model performance. AI4Life aims to incorporate community-driven validation strategies into its workflows.

Integration with the BioImage Archive

The session explored how AI models can be better integrated with the BioImage Archive, focusing on:

  • Programmatic submission of datasets to simplify data sharing for frequent users.
  • Improved interoperability between the BioImage Archive and repositories like EMPIAR, especially for cryo-electron tomography (cryo-ET) data.
  • Standardization of data formats and metadata to enhance AI-readiness and reusability.

Related events and community engagement

The AI4Life team and the Euro-BioImaging community continue to foster dialogue through upcoming events:

Resources

Participants and interested researchers can explore AI4Life’s resources:

Slides

(Pictures by Ayoub El Ghadraoui)

Categories
News

AI4Life highlighted as a European Commission Success Story

AI4Life highlighted as a European Commission Success Story

The European Commission (EC) Directorate-General for Research and Innovation (DG RTD) has recognised the AI4Life project for its achievements, selecting it as a success story on the DG RTD website. This highlights AI4Life’s relevance and significant impact on AI-powered bioimage analysis in Europe.

Our efforts to make AI-powered bioimage analysis accessible to all, through open calls supporting scientists to community-driven initiatives, have been recognised with this milestone.

This acknowledgement reinforces AI4Life’s role in breaking down barriers to AI adoption and fostering collaboration with initiatives like AI4EOSC to strengthen the EOSC research ecosystem.

Read the full story here: https://projects.research-and-innovation.ec.europa.eu/en/projects/success-stories/all/making-ai-accessible-life-science-research

 

Want to connect with the AI4Life community? Join our upcoming community event.

Categories
News

Third AI4Life Open Call: Announcement of selected projects

Third AI4Life Open Call: 
Announcement of selected projects

by Florian Jug & Beatriz Serrano-Solano

The third AI4Life open call received 28 applications from diverse scientific disciplines.

We are again distinguishing two phases, a consultancy phase, and a project execution phase. This is to maximize the value of our service to the scientific community by providing feedback (e.g. feasibility, existing tools, approaches we would take) to a larger number of projects, and then taking on a small number of projects for more in-depth support based on informed opinions of experts regarding the readiness of the project (as we also did last year).

The selection of projects, for both phases, was conducted by an international group of project reviewers, similar to the procedure we conducted in OC2.

Projects selected for Phase 1: Project Consultation

1 – Mechanism of action of insecticidal proteins produced by Bacillus thuringiensis bacteria

3 – Improving nuclei segmentation using Cellprofiler and StarDist

10 – Feeding or Fading? Automated image analysis of white corals’ behavior to unlock natural patterns, stress responses, and conservation needs responses, and conservation needs

11 – 3D Matrix Motility Map (3DM³)

14 – The speed of life in trees — linking wood anatomy with wood lifespan and tree growth

17 – Imprints of Wind Disturbances  on Wood Anatomy

18 – Detection of Nuclear Pore Complexes with shape variability imaged by DNA PAINT

19 – Segmentation of sparse bacteria in human tissue

23 – Smart electron tomography data acquisition

24 – Determine the interaction between neutrophils and preneoplastic cells in a zebrafish skin tumour initiation model

25 – Scalable high-dimensional subcellular profiling in 3D

26 – Automatic microtubule doublet picking in tomograms

27 – 3D cell type identification in lung organoids

Please note that projects with IDs 1 and 24 did not respond to our invitation for consultation even after multiple reminders. Hence, we conducted consultations on a total of 10 projects.

Consultation sessions were typically conducted with between three to five AI4Life experts and consisted of a one-hour discussion during which the experts could ask additional questions regarding the project and then provide guidance regarding the feasibility of the respective projects. Along with this kind of “reality check”, the experts also aimed at giving tangible advice on existing tools and methods that might help the consulted researchers in improving their workflows and achieving the desired analysis results.

Projects selected for Phase 2: Deep Learning project support

With the additional information from the consultancy phase at hand, the group of AI4Life experts came together and discussed the most promising projects to be taken on in Phase 2 (deep-learning support). As last year, this decision was guided by a combination of project feasibility, readiness, re-usability by the community, and how well each project fits within the experiences and capabilities of the AI4Life team itself. These considerations were combined with the intention to maximize the utility and impact on the bioimage analysis community of this Open Call.

3 – Improving nuclei segmentation using Cellprofiler and StarDist

10 – Feeding or Fading? Automated image analysis of white corals’ behavior to unlock natural patterns, stress responses, and conservation needs responses, and conservation needs

11 – 3D Matrix Motility Map (3DM³)

14 – The speed of life in trees — linking wood anatomy with wood lifespan and tree growth

17 – Imprints of Wind Disturbances  on Wood Anatomy

18 – Detection of Nuclear Pore Complexes with shape variability imaged by DNA PAINT

19 – Segmentation of sparse bacteria in human tissue

26 – Automatic microtubule doublet picking in tomograms

Who was involved in the process?

In contrast to last year, we conducted the pre-screening and selection for the consultancy phase within AI4Life and did not include external reviewers.

If you missed our open calls and would like to benefit from a similar service, we recommend Euro-BioImaging’s Web Portal, where you can access a network of experts in the field of image analysis. You can also find a bioimage analyst in your geographic and scientific area in the GloBIAS database of bioimage analysts. Please note that this service may involve associated costs, but access funds for certain research topics are available through some initiatives. We and Euro-Bioimaging will be happy to help you find suitable opportunities!

Categories
News

AI4Life: Advancing AI Adoption in Life Sciences

AI4Life: Advancing AI Adoption in Life Sciences

AI4Life has been working alongside Life Science Research Infrastructures (LS RIs) to address the growing need for AI-driven bioimage analysis. Through a combination of outreach, community engagement, hackathons, and strategic collaborations, the project has significantly expanded access to AI tools for researchers worldwide.

Key achievements

Outreach and Dissemination

AI4Life partners and collaborators have engaged in extensive dissemination activities (175 by mid-February 2025) to promote the project’s services and opportunities, reaching diverse audiences across multiple countries and scientific communities. These activities have taken various forms, including conferences, training events, and scientific meetings, ensuring broad engagement with researchers, industry professionals, and the wider life sciences community. 

The activity types were mostly scientific conferences, followed closely by education/training events (Figure 1).

Conferences provided AI4Life with a platform to showcase the project through presentations and posters. At the same time, training initiatives played an essential role in building community around the BioImage Archive (BIA) and the BioImage Model Zoo (BMZ). Additionally, the project established collaborations with EU-funded projects and other scientific initiatives to strengthen AI4Life’s impact and sustainability.

The geographical distribution of the dissemination activities (Figure 2) highlights a strong international presence even beyond the European scope, together with numerous online events, which enabled global reach and accessibility.

Figure 1. Outreach and dissemination activities from the beginning of the project in September 2022 until February 2025 by type of activity.

Figure 2. Outreach and dissemination activities from the beginning of the project in September 2022 until February 2025 by location.

Community Engagement

AI4Life has organised and contributed to multiple workshops and training events to introduce AI methods to the life science community. These events often included presentations and hands-on sessions by AI4Life experts.

As part of its outreach efforts, AI4Life has collaborated closely with Life Science Research Infrastructures (LS RIs) through Work Package 7, engaging with partners at Euro-BioImaging, EU-OPENSCREEN, the European Marine Biological Resource Centre (EMBRC), EMPHASIS, and Instruct-ERIC. Through monthly meetings, AI4Life has raised awareness among the partner LS RIs about the opportunities it offers to their scientific communities while gaining insight into their specific needs. This ongoing dialogue has enabled AI4Life to contribute meaningfully to LS RI-led initiatives, ensuring its presence at relevant events. A key example is the EU-OPENSCREEN Autumn Training School, where AI4Life members showcased the offered services to demonstrate their impact on bioimage analysis.

Hackathons

Hackathons have been essential to AI4Life, bringing together project members and key stakeholders to develop new features and enhance existing services (Figure 3).

These events also fostered collaboration by inviting industry partners, community partners and participants from other initiatives to contribute their knowledge. Community partners typically include organisations, companies, research groups, or software teams that can consume and/or produce resources of the BioImage Model Zoo. By involving them in shared development activities, AI4Life ensures interoperability across resources.

We strategically aligned these events with General Assemblies to maximise impact and efficiency, leveraging attendees’ expertise while minimising travel requirements.

Figure 3. Some of the Hackathons organised by AI4Life.

Strategic Collaborations and Recognition

AI4Life has strategically partnered with other EU-funded projects to increase the project’s sustainability and impact. The European Commission (EC) Directorate-General (DG) for Research and Innovation has acknowledged the project’s achievements and selected AI4Life to be showcased at the European Research and Innovation Days 2024 and as a success story on the DG Research and Innovation website.

As AI4Life moves into 2025, the project remains committed to expanding its outreach, further developing AI tools, and continuing international collaborations. By maintaining strong engagement with the life sciences community, AI4Life will ensure that its contributions to bioimage analysis remain at the forefront of scientific innovation.

For more information on AI4Life’s activities, read the deliverable D7.2 available in Zenodo and for upcoming events, visit the events section: https://ai4life.eurobioimaging.eu/events.

Categories
News

Expanding AI4Life training resources: New and updated materials

Expanding AI4Life Training Resources: New and Updated Materials

by Caterina Fuster-Barceló

To enhance accessibility and usability, the European consortium AI4Life has expanded and updated the BioImage Model Zoo (BMZ, BioImage.IO) training materials as part of the D3.3 Deliverable (available in zenodo). This effort involved evaluating existing resources, introducing new materials, and updating documentation to ensure comprehensive guidance for both developers and end-users of the BMZ.

Currently, the AI4Life training material list includes over 30 different training materials, spanning website documentation, case studies, video tutorials, computational notebooks, and slides. The majority of resources are targeted at end-users (47.1%) and developers (32.4%), with additional materials designed to serve both audiences (20.6%). The updated distribution of training materials and their target audiences is visualized in the figures below.

Key Insights from the Updated Training Materials

  • Video tutorials (14) remain the most used format, followed by website documentation (9) and computational notebooks (6).
  • Nearly half of the training materials cater to end-users, helping them navigate BMZ and integrate AI models into workflows.

Training Documentation Aligned with the AI Model Lifecycle

To ensure that documentation is structured in a way that is intuitive and practical for users, we have aligned its training materials with the lifecycle of a Deep Learning (DL) model in the BMZ. The image below illustrates this workflow-based documentation approach, categorizing resources according to the key stages of models and the targeted users of each case.

Each stage of the DL model lifecycle—model development, packaging, upload, accessibility, deployment, and fine-tuning—is supported by a combination of website documentation, video tutorials, computational notebooks, and case studies. Developer-focused documentation (highlighted in orange) ensures that DL models are properly packaged, documented, and submitted to BMZ, while end-user documentation (highlighted in blue) provides guidance on how to access and integrate models into image analysis workflows.

Highlighted Updates and New Materials

1. New Guided Tutorials on YouTube. AI4Life has added three new step-by-step video tutorials to help users and developers navigate BMZ more efficiently.

2. Comprehensive Guide on Running BMZ Models in Community Partner Tools. A new documentation resource provides clear, step-by-step instructions on how to execute BMZ models across different bioimage analysis software.

3. Expanded Documentation on the Resource Description File (RDF). To improve model documentation clarity, AI4Life has introduced updated materials explaining the BMZ Resource Description File (RDF) in various formats, making it more accessible to different user groups.

These updates strengthen AI4Life’s commitment to providing high-quality, accessible training materials that support the growth and usability of the BioImage Model Zoo. By continuously refining and expanding these resources, AI4Life ensures that bioimage analysis researchers and developers have the tools they need to efficiently integrate, deploy, and utilize AI models in their scientific workflows.

Categories
News

AI4Life and AI4EOSC join forces to facilitate AI model deployment for Bioimage Analysis

AI4Life and AI4EOSC join forces to facilitate AI model deployment for Bioimage Analysis

AI4Life and AI4EOSC started an effective collaboration at the AI4Life workshop in Madrid, where AI4EOSC members participated and introduced the AI4OS software stack and the AI platform, which enables training, deployment, sharing and monitoring of AI models. Building on this partnership, AI4EOSC has developed a tailored version of the AI platform specifically for AI4Life, enabling researchers to deploy and run AI models via the AI4Life Cloud Marketplace (Figure 1).

Figure 1. AI4Life Cloud Marketplace accessible at  https://ai4life.cloud.ai4eosc.eu/marketplace

This new platform simplifies the process of launching AI models for bioimaging tasks by integrating with AI4EOSC’s computational infrastructure. Researchers can now take advantage of this scalable and user-friendly environment to run models from the BioImage Model Zoo, enhancing accessibility and reproducibility in bioimage analysis (Figures 2 and 3).  

Figure 2. BioImage Model Zoo models imported to the AI4Life Cloud Marketplace

Figure 3. Example model from BioImage Model Zoo

Currently, the system supports models adhering to the v5 of the BioImage.IO specifications. This serves as a strong example of how FAIR, standardized models pave the way for interoperability, making resources more accessible across different platforms. The success of this integration was highlighted at the EOSC Winter School in Sevilla, where it was showcased as a model case for interoperability between research infrastructures.  

For detailed guidance on deploying models via this platform, refer to the official AI4EOSC dashboard documentation:

https://docs.ai4eosc.eu/en/latest/howtos/deploy/external.html  

This collaboration reinforces AI4Life’s mission to make AI-powered bioimage analysis more accessible, interoperable, and integrated into the European Open Science Cloud (EOSC). Stay tuned for further updates as we continue to advance AI capabilities for the bioimaging community!

Categories
News

New Community Partner: CAREamics!

New Community Partner: CAREamics!

by Joran Deschamps, Beatriz Serrano-Solano

CAREamics, a library that simplifies using deep-learning denoising algorithms, has joined the BioImage Model Zoo.

With CAREamics, users can improve their analysis pipelines by removing noise from their microscopy images through robust image restoration. By leveraging popular deep learning methods, CAREamics empowers researchers to recover high-quality data from low light or fast imaging,  ensuring lower sample phototoxicity, better quantification, and reproducibility in their experiments.

As a Community Partner, CAREamics now provides ready-to-use models compatible with the BioImage.io model format. 

Discover CAREamics models on the BioImage Model Zoo: https://bioimage.io/#/?partner=careamics&type=application