DMEA 2026

Trade exhibition and congress for digital healthcare / April 21, 2026 - April 23, 2026

Connecting Digital Health

DMEA is Europe's leading event for digital health, bringing together experts from healthcare, IT, and research to explore the future of healthcare technology. The upcoming DMEA 2026 will take place from April 21 to 23, 2026, at Messe Berlin in Germany. The event features keynote speeches, panel discussions, and exhibitions showcasing the latest innovations in healthcare IT, artificial intelligence, telemedicine, and digital patient care. It serves as a central platform for networking, knowledge exchange, and collaboration, driving the digital transformation of healthcare.

Experts of the Fraunhofer Institute for Digital Medicine MEVIS are looking forward to welcoming you at the joint booth of the Fraunhofer-Gesellschaft (Hall 2.2 | Booth D-109).

At this year's DMEA, Fraunhofer MEVIS presents various solutions for digital medicine:

Efficient AI Training and Curation of Medical Image Collections with CuraMate
ExaminAI: Fine-Grained AI Performance Monitoring and Failure Analysis
Mitral Valve Analysis and Therapy Planning Tool
Assistant LUKAS for Advanced Patient Care and Nursing Efficiency
Interactive Courses for Medical Training with CuraMate

Event details

Location

Messe Berlin
Berlin ExpoCenter City | Messe-Süd
(Hall 2.2 | Booth D-109)
Jafféstraße, 14055 Berlin

Date

April 21, 2026 - April 23, 2026

Download this event as iCal

Efficient AI Training and Curation of Medical Image Collections with CuraMate

Screenshot of local correction of an erroneous part of a spleen segmentation. The model prediction (blue) is erroneous and uncertain (indicated by red areas) because the initial model was only trained with two labeled cases. Subsequent models learn from the user's corrections and perform better. — © Fraunhofer MEVIS

CuraMate AI Development is a versatile, customisable software platform for collaborative research and development of AI models. It facilitates the tedious process of annotating medical image data, with excellent support for DICOM and other formats, CT, MRI, ultrasound, or video data. Our uncertainty-driven training loop minimises annotation time and maximises model accuracy through an efficient, iterative process. Starting with just a few training cases, the model generates initial segmentations and computes areas of uncertainty at the same time. These highlighted areas guide users directly to cases and regions needing correction.

ExaminAI: Fine-Grained AI Performance Monitoring and Failure Analysis

Data quality and AI model quality assessment — © Fraunhofer MEVIS

Medical AI models can fail silently — not uniformly, but in specific, identifiable subgroups. Shortcut learning, imaging artifacts, patient implants, or labeling errors may each affect performance in distinct, hidden ways that aggregate metrics simply cannot reveal.
Our software combines visual clustering of volumetric MRI and CT datasets with rigorous stratified performance analysis. Embedding-based clustering surfaces similar images automatically, enabling rapid visual inspection of data quality issues and model failure patterns across hundreds or thousands of cases. Carefully chosen performance metrics are then computed per cluster and stratified by clinical and technical covariates — including demographics, scanner parameters, and acquisition protocols — with statistical rigor.
The result: a clear and actionable interactive picture of where a model works, where it struggles, and why — empowering developers and clinical teams to build more robust, trustworthy AI.

Mitral Valve Analysis and Therapy Planning Tool

We present a software solution that generates patient-specific digital twins of the left heart from computed tomography scans to support planning of minimally invasive interventions. Users can review and refine automatically generated segmentation masks with intuitive tools. Anatomical landmarks describing features of the heart valve are extracted and combined into a simulation-ready heart model. From this digital twin, key quantitative parameters describing the valve functionality are derived to support objective assessment. Based on the digital twin we can simulate valve closure and the effects of minimally invasive interventions. This allows clinicians to evaluate different therapy options before the intervention, e.g., different devices and implantation positions.

Precise Planning for Heart Surgery

Assistant LUKAS for Advanced Patient Care and Nursing Efficiency

LUKAS is an assistance system for nursing wards in hospitals that is used by both patients and nursing staff via voice and touch control. It is designed to improve the quality of stay for patients, to provide comprehensive information and to simplify documentation requirements and the coordination of care processes.
A unique feature of LUKAS is its ability to obtain information from a variety of sources and evaluate it independently. LUKAS draws on data from the clinical information system, the nursing documentation system and various sensors such as sleeping mats and wearable devices.
During the working day, LUKAS carries out documentation tasks as automatically as possible and can go through open points with the nursing staff at the end of the shift. This significantly reduces the manual documentation effort and helps to avoid gaps in documentation.
In situations when quick communication is required, LUKAS can proactively contact both patients and nursing staff making him a reliable partner in everyday nursing.

SmarthHospital.NRW/KI-gestützte Demonstratoren (german)

Interactive Courses for Medical Training with CuraMate

Interactive medical education using the example of cardiovascular MRI images — © Fraunhofer MEVIS

Our interactive medical education platform delivers advanced skills training through hands-on annotation, segmentation, and quantification of multimodal medical images. By simulating real clinical workstations, the platform delivers authentic learning experiences and assessment capabilities that extend beyond conventional questionnaires. The system integrates seamlessly with existing educational solutions, offers configurable workflows, and provides cloud-based access for scalable, browser-ready deployment. Visit us at DMEA to see our example implementation for cardiovascular imaging education in action and discover its potential for different medical imaging specialities.

CuraMate Education