EGI2025 Day 2 - Addressing AI needs for processing of imaging data скачать в хорошем качестве

EGI2025 Day 2 - Addressing AI needs for processing of imaging data

EGI2025 - Addressing AI needs for processing of imaging data: the case of Aquatic Science and the iMagine projectTechnical advancements and cost drops in sensor technologies have radically broadened scientific image and video production. The scientific world now requires scaling up its capacity to analyse the vast amount of images produced. While Artificial Intelligence (AI) is a promising solution for automated image analysis, the availability of computational resources, AI tools and AI expertise are serious limiting factors to analysis. Although EuroHPC (e.g., AI factories) and other centralised HPC facilities provide infrastructure for AI applications, the problem requires a more democratised approach that expands the applicability of AI-based image analysis through national and institutional compute and data resources and distributed support teams. The iMagine project is an EU-funded initiative, coordinated by the EGI Foundation, that provides answers to these challenges and democratises AI-based image analysis within aquatic sciences. iMagine nears its conclusion in August 2025, and with this plenary session, it offers a forward-looking reflection on its contributions to the field. The session will highlight the project's achievements: (1) the iMagine AI platform, (2) the Competence Centre, (3) the federated cloud infrastructure, (4) labelled image sets, (5) thematic image analysis applications, (6) training and outreach programme, and will provide a scientific-technical deep-dive through a live demonstration of the FlowCam Phytoplankton Identification service. Following the demonstration, an expert panel will explore critical topics for future work, including: lessons learned from iMagine to inform future initiatives; the adaptability of the iMagine AI tools to other scientific domains; and strategies to ensure broad and long-term uptake of AI through knowledge transfer, training, and community engagement. The interactive session will invite participants to reflect on the future of AI in science, building on the experience and outcomes of iMagine. The conversation aims to spark insights, share experiences, and identify actionable paths for sustaining and expanding the innovations developed within iMagine. Facilitator: Smitesh Jain (EGI Foundation, iMagine Innovation Manager) 1. 15’ AI-based image analysis for aquatic sciences - Needs and approaches. Presenter: Gergely Sipos (EGI Foundation, iMagine Project Coordinator) 2. 20’ Challenges and solutions - Collaboration & Technology. Presenter: Fahimeh Alibabaei KIT, Álvaro López García (CSIC) 3. 20’ A featured example with a demo: FlowCam phytoplankton identification. Presenter: Wout Decrop (VLIZ) 4. 15’ Achievements and next steps - Ilaria Fava (EGI Foundation, iMagine Communication Manager) 5. 35’ Panel discussion and Q&A