AI platform makes microscopy image analysis more accessible

An international research collaboration has developed a platform that empowers life scientists to harness deep learning techniques for biomedical research. The platform is called DL4MicEverywhere and makes artificial intelligence (AI) accessible for analysing microscopy images, empowering researchers regardless of their computational expertise. It has been described in the journal Nature Methods.

Deep learning, a subset of AI, has transformed the analysis of large and complex microscopy datasets, enabling automatic identification, tracking and analysis of cells and subcellular structures. Despite these advancements, the need for computing resources and AI expertise has limited the adoption of these techniques in life science research.

DL4MicEverywhere addresses these challenges by offering an intuitive interface that allows researchers to train and apply deep learning models on various computing infrastructures, from laptops to high-performance clusters. Its development was made possible through a collaboration of experts in computer science, bioimage analysis and microscopy, co-led by Professor Ricardo Henriques’s laboratory at the Instituto Gulbenkian de Ciências (IGC) and Professor Guillaume Jacquemet’s laboratory at Åbo Akademi University, with crucial contributions from the AI4Life consortium.

“DL4MicEverywhere establishes a bridge between AI technological advances and biomedical research,” said first author Ivan Hidalgo-Cenamor, a researcher at IGC. “With it, researchers gain access to cutting-edge methods, enabling them to automatically analyse their microscopy data and potentially discover new biological insights.”

DL4MicEverywhere builds on the team’s previous work, ZeroCostDL4Mic, introducing several key advancements. It facilitates the training and deployment of models across different computational environments by encapsulating deep learning workflows in shareable and reproducible Docker containers. The platform also features a user-friendly graphical interface and expands the collection of available models for common microscopy image analysis tasks. It will be available as an open-source resource at https://github.com/HenriquesLab/DL4MicEverywhere, lowering the barriers to advanced microscopy image analysis and hopefully enabling breakthroughs in fields ranging from basic cell biology to drug discovery and personalised medicine.

“DL4MicEverywhere aims to democratise AI for microscopy by promoting community contributions and adhering to FAIR principles — making models findable, accessible, interoperable and reusable,” said IGC’s Dr Estibaliz Gómez-de-Mariscal and Åbo Akademi University’s Dr Joanna Pylvänäinen.

“It will allow life scientists without coding experience to use deep learning on large numbers of microscopy images and videos to make discoveries. This will revolutionise how researchers plan their experiments and extract new information from microscopy datasets.”

Henriques and Jacquemet agreed that the work represents an important milestone in making AI more accessible and reusable for the microscopy community.

“By enabling researchers to share their models and analysis pipelines easily, we can accelerate discoveries and enhance reproducibility in biomedical research,” they said. “DL4MicEverywhere has the potential to be transformative for the life sciences; it aligns with our vision in AI4Life to develop sustainable AI solutions that empower researchers and drive innovation in health care and beyond.”

Image credit: iStock.com/AlexRaths