Skip to content
Home » SEM-CMR

SEM-CMR

    Our Work

    SEM-CMR

    Deep semantic volumetric reconstruction of cardiac magnetic resonance imaging

    Starting date

    18 May 2026

    End date

    17 May 2027


    SEM-CMR

    Deep semantic volumetric reconstruction of cardiac magnetic resonance imaging

    SEM-CMR aims to advance cardiac magnetic resonance (CMR) imaging by developing artificial intelligence methods capable of reconstructing 3D images of myocardial perfusion from routine clinical scans. Coronary artery disease remains the leading cause of death worldwide, yet current CMR perfusion imaging captures only a limited portion of the heart, potentially overlooking clinically important regions of reduced blood flow. SEM-CMR addresses this challenge by combining patient-specific anatomical information with semantic deep learning to generate physiologically meaningful whole-heart perfusion reconstructions, validated against invasive reference techniques. Beyond improving the detection of myocardial ischaemia, the reconstructed 3D heart provides a tailored anatomical reference frame for each patient, enabling the co-registration and joint analysis of complementary CMR biomarkers, such as perfusion, function and tissue characterization, paving the way towards more comprehensive and personalised cardiac assessment. Using retrospective clinical datasets, the project will establish the feasibility of this approach while laying the scientific and methodological foundations for a future European Research Council (ERC) Consolidator Grant in AI-driven, physiology-aware cardiac imaging. The project is funded by the Republic of Cyprus, through the Research and Innovation Foundation.

    Contact Person:

    Dr Xenios Milidonis

    Email: x.milidonis@cyens.org.cy

    ↑