AI4Life

AI4Life_Logo
Menu

Month: October 2025

Categories
News

In retrospect: AI4Life key results and impact

In retrospect: AI4Life key results and impact

The overall aim of the Horizon Europe-funded AI4Life project (September 2022 – August 2025) was to bridge the gap between life sciences and artificial intelligence (AI), focusing on bioimage analysis. Coordinated by Euro-BioImaging ERIC and supported by other leading European research infrastructures, the project aimed to democratise access to AI-based image analysis tools, foster interoperability, and promote FAIR and Open Science principles.

Over three years, AI4Life delivered a robust infrastructure, community-driven services, and standards that enable researchers to share, discover, and reuse AI models and data for bioimage analysis across disciplines, thus promoting sustainable AI adoption in life science research.

The BioImage Model Zoo (BMZ) is at the heart of these achievements, a centralised platform that democratises access to pre-trained AI models. By offering models with standardised metadata, reproducible deployment options, and integration with widely used analysis tools, the BMZ empowers life scientists – regardless of their computational expertise – to apply cutting-edge AI methods to their imaging data.

Complementing the BMZ, AI4Life developed the BioEngine, a scalable service for testing and deploying models and Hypha, a modular platform now enhanced with AI agent support for automated data management and scientific discovery. These infrastructures reduced the technical overhead traditionally associated with AI adoption.

To further ease model discovery and use, the BioImage.IO Chatbot was introduced, allowing users to explore and apply models interactively. AI4Life also launched a series of Open Calls and Public Challenges, directly supporting life scientists and generating reusable workflows, datasets, and benchmarking tasks. These initiatives fostered community engagement and accelerated the development of robust, generalizable AI solutions.

The project’s commitment to FAIR principles is evident in its data services, which include curated datasets with ground-truth annotations, standards for AI-ready data, and deposition services to ensure reproducibility. These resources are openly available via the BioImage Archive and Zenodo, reinforcing transparency and reuse.

Crucial for all these efforts was the AI4Life community engagement and building. From the outset, AI4Life strongly emphasised developing its resources with the community, not just for it. The project created a network of community partners, including academic developers, research infrastructures, and software tool creators who contributed compatible models, datasets, and technical expertise. This co-development approach ensured that tools and standards reflected real user needs and that outcomes could be immediately integrated into existing research practices.

Altogether, these results empower researchers, developers, and infrastructures to adopt and contribute AI solutions with minimal technical barriers. Sustainability is ensured through integration with Euro-BioImaging, EMBL-EBI (BioImage Archive), EOSC platforms, and a newly established AI4Life governance framework involving founding partners and community boards.

From results to impact

Through its open outputs, collaborations, and measurable uptake within the research community, AI4Life has generated meaningful scientific impact. Over the project, more than 40 scientific publications – including 27 peer-reviewed articles – were produced, accumulating nearly 200 citations. All deliverables, datasets, and software are available via Zenodo, the BMZ, and the BioImage Archive, demonstrating the project’s commitment to transparency and early dissemination.

AI4Life has also made a lasting contribution to improving interoperability across widely used image analysis tools. Researchers can now apply advanced deep-learning models without extensive technical expertise by enabling direct access to BMZ models within Fiji, ilastik, ImJoy, CAREamics, and BiaPy.

Training and outreach activities have also been central to AI4Life’s scientific impact. Through workshops, hackathons, and hands-on courses across Europe, AI4Life provided targeted training for computational and experimental scientists. The Open Calls and Public Challenges created collaborative opportunities that yielded new workflows and ground-truth datasets, some of which have already resulted in publications and new benchmark resources for method development.

These efforts have built an active interdisciplinary community that connects developers and users, enabling AI-based analysis to become a regular component of biological research rather than a specialist activity.

AI4Life has also significantly contributed to the European and international Open Science and FAIR data agendas. The project promoted complete transparency in research by making all its outputs – data, models, and code – publicly accessible under open licences. AI4Life operationalised key components of the EU’s Open Science and data management policies by embedding reproducibility, open licensing, and persistent identifiers into all its services. These practices make the project a practical example of how FAIR and open principles can be implemented in everyday scientific workflows. 

AI4Life’s standardisation efforts represented key policy contributions. The MIFA (Metadata, Incentives, Formats, Accessibility) recommendations developed in AI4Life and the community have become an influential reference for the FAIR sharing of annotated image datasets. The project also advanced the BioImage.IO model specification and aligned its work with the Research Data Alliance FAIR4ML Interest Group, contributing to broader standards for machine-learning metadata. These efforts provide a concrete implementation of FAIR principles for AI models and datasets already being adopted by community software tools.

Integration into the European Open Science Cloud (EOSC) further ensured AI4Life’s policy, scientific relevance, and long-term availability. Through collaboration with the AI4EOSC project, AI4Life models are now deployable within the AI4OS platform, connecting life science users directly with European computing resources.

Beyond its scientific and policy contributions, AI4Life has generated early economic, societal, and environmental benefits.

Economically, the infrastructure and services developed within AI4Life reduce duplication of effort and support cost efficiency. By offering standardised, validated AI models and interoperable tools, laboratories and companies can avoid redeveloping similar solutions and instead adapt existing ones. Collaborations with industry leaders such as Leica Microsystems have demonstrated how project outputs can be incorporated into commercial image analysis platforms, expanding their reach to a global user base. Moreover, the creation of the KTH spin-off Amun AI, which builds on AI4Life’s Hypha software improvements, illustrates how project-generated technologies can evolve into business applications.

AI4Life’s approach also has environmental implications, as the reuse of models reduces redundant computation and data storage. Adopting shared infrastructures and containerised deployments promotes more resource-efficient research practices, aligning with broader European goals for digital sustainability.

From a societal perspective, AI4Life has contributed to responsible and transparent AI adoption in life sciences. The project enhances trust and reproducibility in computational AI-based research by linking models to peer-reviewed publications and training data and enforcing open licensing.

 

A sustainable foundation for open and impactful bioimage analysis

AI4Life has laid the foundation for an active, sustainable ecosystem of AI-powered bioimage analysis. Through its infrastructure, standards, and community engagement, AI4Life has created a robust framework that strengthens Europe’s capacity for data-driven discovery, ensuring that tools are and remain FAIR, open, and impactful, and supporting the responsible and collaborative use of AI in the life sciences.

Stay connected

Community governance is open and transparent. Anyone can join our weekly Wednesday meetings at 16:00 CE(S)T:

  • First week of the month: Community Board (with current and prospective Community Partners).
  • Other weeks: Technical Board (with Task Forces spun up as needed).

The same Zoom link will be used for all meetings, and agendas are kept in HackMD, and meeting notes are posted to GitHub for wider visibility (details here).

Follow Euro-BioImaging and our partners to hear more about upcoming developments, community opportunities, and how you can get involved in shaping the BioImage Model Zoo. 

And don’t forget to subscribe to the Euro-BioImaging Newsletter!

Linkedin Instagram Link Youtube Twitter
Categories
News

In retrospect: How AI4Life collaborated with Industry

In retrospect: How AI4Life collaborated with Industry

Interactions and exchanges with industry have been important aspects of the AI4Life project. By actively engaging with stakeholders across microscopy, pharmaceuticals, and technology, AI4Life ensured that its tools, standards, and infrastructure were not only scientifically robust but also relevant and usable in real-world industrial environments.

Engaging industry from the start

Throughout its project lifetime, AI4Life maintained a proactive approach to industry collaboration. The project held dedicated meetings with key stakeholders – including, for instance, the Euro-BioImaging Industry Board, Bayer, Zeiss, and Leica – to exchange ideas, align on technical needs, and explore integration opportunities. Company representatives also participated in AI4Life hackathons, contributing to open discussions and co-development efforts with academic and infrastructure partners. This open-door strategy helped foster trust and mutual understanding between sectors.

Bringing AI models into commercial microscopy platforms

One of the most tangible outcomes of AI4Life’s industry engagement was its collaboration with Leica Microsystems. Together, they successfully integrated deep-learning models from the BioImage Model Zoo (BMZ) into Leica’s AIVIA software, a commercial AI-powered image analysis platform. Leica converted BMZ models into its own repository format, enabling interoperability with AIVIA. This milestone demonstrated that AI4Life’s standards and infrastructure were mature enough for commercial deployment. Updates from this collaboration were shared openly on the AI4Life website.

ZEISS also broadened its collaboration with AI4Life. Their Arivis platform started supporting pre-trained AI models, custom-trained models, and external technologies such as CAREamics, an AI4Life community partner. Since CAREamics models follow the AI4Life model specifications, this opens a new path for BMZ models to be used directly within commercial workflows.

Supporting innovation through Open Calls

AI4Life’s Open Calls for AI-based image analysis support were open to both academic and industry participants, reinforcing the project’s inclusive philosophy. Among the selected projects was one from RD Néphrologie, a French SME, which received support to study the impact of Chronic Kidney Disease on collagen density in tissues such as the heart and kidney. This example highlighted how AI4Life’s infrastructure could be applied to pressing biomedical challenges in both academic and commercial contexts.

Working with technology leaders

AI4Life partners also participated in NVIDIA’s Early Access Program to test the Triton Inference Server. These evaluations provided valuable insights into the scalability of the BioEngine model testing service and informed improvements to the platform’s technical performance.

Supporting sustainability through innovation

The Hypha platform, the core component used and refined in AI4Life for serving AI models, became the foundation for the spin-off company Amun AI. Amun AI’s plug-and-play version of Hypha builds on AI4Life technology, includes confidential computing, and is designed for large-scale data management and AI models serving research institutes and industrial users. This transition from research prototype to commercial product exemplifies AI4Life’s commitment to sustainable innovation and long-term impact.

Why it matters

These collaborations contributed to the early development of AI adoption in bioimaging within industry. By aligning technical development with industry needs, enabling model interoperability, supporting SMEs through Open Calls, testing AI models in commercial environments, and promoting spin-offs, AI4Life laid the foundation for an active ecosystem and long-term impact. Its open, inclusive, and standards-driven approach ensures that the tools created during the project will continue to benefit both academic and industrial users well beyond the official project conclusion.

Stay connected

Community governance is open and transparent. Anyone can join our weekly Wednesday meetings at 16:00 CE(S)T:

  • First week of the month: Community Board (with current and prospective Community Partners).
  • Other weeks: Technical Board (with Task Forces spun up as needed).

The same Zoom link will be used for all meetings, and agendas are kept in HackMD, and meeting notes are posted to GitHub for wider visibility (details here).

Follow Euro-BioImaging and our partners to hear more about upcoming developments, community opportunities, and how you can get involved in shaping the BioImage Model Zoo. 

And don’t forget to subscribe to the Euro-BioImaging Newsletter!

Linkedin Instagram Link Youtube Twitter
Categories
News

In retrospect: Open Science in AI4Life

In retrospect: Open Science in AI4Life

From the beginning, AI4Life has treated Open Science as a foundation for how research should be shared, reused, and built upon. It has ensured that its results are accessible to everyone, not just those inside the consortium.

Sharing early, sharing widely

To make knowledge travel faster, AI4Life partners share their manuscripts and technical reports as preprints long before formal publication on platforms like bioRxiv, arXiv, and Zenodo. This way, project partners have invited feedback from the global community and prevented valuable work from staying behind paywalls or review delays.

AI4Life also opened its infrastructure and services to the outside world through Open Calls and Challenges, allowing researchers and developers everywhere to both contribute to and benefit from solutions to real imaging problems.

Open tools, open infrastructure

A significant contribution of AI4Life to Open Science is the development of tools and platforms that make AI models for image analysis easy to find, run, and reuse:

  • BioImage Model Zoo: a public repository of AI models for microscopy and bioimaging, where each model is openly licensed, versioned, and citable.

  • BioEngine: a web-based environment that lets users run AI models and workflows in the browser or on cloud resources.

  • Integration with community tools: Platforms like DL4MicEverywhere, BiaPy, CAREamics, deepImageJ, ilastik, ImJoy, SpotMAX, and ZeroCostDL4Mic bring these AI4Life tools directly into the daily workflows of life scientists.

Open by design

AI4Life ensures openness by:

  • Publishing data, code, and models under open licenses (MIT, BSD, Creative Commons).
  • Assigning DOIs and permanent identifiers so each model version can be cited and traced.
  • Storing resources openly in Zenodo, the BioImage Model Zoo (BioImage.IO), and the BioImage Archive.
  • Requiring reproducibility tests for every model in the BioImage Model Zoo, models must include test inputs/outputs and pass automated checks.
  • Using containerised environments (e.g. DL4MicEverywhere) allows tools to run the same way on any computer, cluster, or cloud system.

A community effort – Working together, not alone

AI4Life has actively fostered open engagement with both the imaging and AI development communities, recognising that building impactful infrastructure requires collective effort. Through workshops, hackathons, and other collaborative activities, the project has invited external researchers, tool developers, and domain experts to co-create solutions, share insights, and shape the direction of the AI4Life services and infrastructure. This inclusive approach ensures that the tools and standards developed are not only technically sound but also aligned with the real-world needs of diverse scientific communities.

To that end, AI4Life also works closely with European Research Infrastructures (Euro-BioImaging ERIC, EMBRC, EU-OPENSCREEN, Instruct ERIC, EMPHASIS) and a growing network of Community Partners. These organisations help embed AI4Life tools into their daily activities and disseminate their results.

By combining early sharing, open infrastructure, community collaboration, and automated reproducibility checks, AI4Life has built trust, transparency, and long-term impact, ensuring that the tools created today remain useful for researchers tomorrow.

Stay connected

Community governance is open and transparent. Anyone can join our weekly Wednesday meetings at 16:00 CE(S)T:

  • First week of the month: Community Board (with current and prospective Community Partners).
  • Other weeks: Technical Board (with Task Forces spun up as needed).

The same Zoom link will be used for all meetings, and agendas are kept in HackMD, and meeting notes are posted to GitHub for wider visibility (details here).

Follow Euro-BioImaging and our partners to hear more about upcoming developments, community opportunities, and how you can get involved in shaping the BioImage Model Zoo. 

And don’t forget to subscribe to the Euro-BioImaging Newsletter!

Linkedin Instagram Link Youtube Twitter
Categories
News

In retrospect: AI4Life Challenges

In retrospect: AI4Life Public Challenges

AI4Life’s mission isn’t just to assist individual projects , it also seeks to benchmark and push the boundaries of AI methods for bioimage analysis. Over three years, the project has organised three public challenges, inviting the computational community to compete on core imaging tasks under shared datasets, metrics, and conditions.

Overview & evolution

First Challenge: unsupervised denoising

AI4Life launched the inaugural challenge focusing on unsupervised denoising of microscopy images. Rather than relying on paired “noisy/clean” image sets, participants were invited to apply algorithms that learn denoising from noisy data alone.

This setup reflects a practical constraint in microscopy: obtaining “clean” reference images is often difficult or impossible. The goal was to reduce noise while preserving delicate image features like edges, textures, and fine detail.

Four datasets were used, with two types of noise (structured and unstructured) represented. https://ai4life-mdc24.grand-challenge.org/ai4life-mdc24/

The 2024 AI4Life Denoising Challenge drew 104 registrants from 27 countries, resulting in 151 submissions across four leaderboards (each corresponding to a dataset/noise type).
Top results included combinations of algorithms like COSDD and N2V, which delivered strong performance across modules. 

Find out more: https://ai4life.eurobioimaging.eu/ai4life-denoising-challenge-2024-results/

Second & Third Challenges: transitioning to supervised denoising

Following the insights from 2024, the 2025 challenge shifts focus to supervised denoising, now combining noisy and clean images to train models. The change allows for more precise performance evaluation and potentially better denoising when ground truth is available. The 2025 edition moved to include supervised denoising (paired noisy/clean) and a specialised calcium-imaging track. 

The MDC25 and CIDC25 result pages (hosted on Grand-Challenge) hold the detailed leaderboards and entries; the CIDC25 platform remains accessible for benchmarking and late submissions:

These two challenges together drew 91 submissions from 13 countries across eight leaderboards, with more details at: 

https://ai4life.eurobioimaging.eu/ai4life-denoising-challenges-2025-results/

 

Why these challenges matter

  • Benchmarking: Standardised challenges let us compare methods fairly, across diverse datasets and noise types.
  • Broad community engagement: By opening up to anyone (not just project partners), AI4Life attracts fresh ideas and cross-pollination from adjacent fields.

From calls to conversations: reflections from our experts

Challenges are not just competitions; each involves substantial coordination, data curation, evaluation, and community outreach. To bring that human side forward, we recorded conversations with several experts who played key roles in designing, running, and evaluating these challenges.

Watch the full video and hear their stories: 

Categories
News

In retrospect: AI4Life Open Calls

In retrospect: AI4Life Open Calls

Over the past 3 years, AI4Life has launched a series of three annual Open Calls to support life scientists facing unmet image analysis needs. 

These calls invited scientists working with biological or microscopy images to propose analysis problems where deep learning makes a difference. Proposals were selected based not just on novelty, but on broader impact, reusability, and feasibility. 

What each Open Call offered
  • Technical support and collaboration: Project experts provided guidance on the development of analysis workflows, data improvement and deep learning solutions.
  • Consultation phase (from Open Call 2 onward): Before full project selection, top applicants engaged in a consultation phase to assess data readiness, existing methods’ suitability, and possible quick wins. In some cases, the consultation itself was enough to resolve issues.
  • Public dissemination: At the end of each project, the developed workflows, some subset of data, and trained models were made publicly available (e.g. via the BioImage Model Zoo or open archives).
  • Potential elevation into public challenges: Projects selected via the Open Calls addressing common needs could be turned into public challenges, inviting more involvement from the computational community to improve or benchmark solutions.
Evolution through three editions

First Open Call (Spring 2023) 

https://ai4life.eurobioimaging.eu/open-calls/first-open-call/ 

  • 72 proposals submitted from various life science domains, from cell and developmental biology to marine and plant research.
  • 67 aimed to improve existing workflows, while fewer proposed new AI-based methods.
  • The lack of annotated (“ground truth”) data was a recurrent bottleneck, revealing a widespread need for annotation support and AI-readiness training.

Second Open Call (Autumn 2023)

https://ai4life.eurobioimaging.eu/open-calls/second-open-call/ 

  • Introduced a two-phase process, beginning with a 1-hour consultation phase to assess project feasibility before full project support.
  • Project consultations increased interaction between applicants and experts, leading to better alignment of expectations and higher-quality collaborations.

Third Open Call (Spring 2024)

https://ai4life.eurobioimaging.eu/open-calls/third-open-call/

  • Received 28 applications across Europe, showing continuing demand despite a more targeted scope.
  • As in OC2, top applications underwent consultation before selected cases advanced to full technical support.

Across all calls, an international review committee has guided project selection by evaluating scientific impact, feasibility, reusability, and alignment with AI4Life’s capabilities.

Use Cases: real stories of impact

All projects selected through the Open Calls are showcased at the Use Cases page, illustrating how AI4Life support can turn complex bioimage tasks into reproducible, shared pipelines.

https://ai4life.eurobioimaging.eu/open-calls/use-cases 

These use cases show the diversity of problems addressed (2D/3D, multi-modal, multi-label) and how the solutions are made open for reuse.

 

From calls to conversations: reflections from our experts

Behind every open call was a dedicated group of experts guiding applicants, shaping proposals, and co-developing solutions. Watch the video below to hear their stories: