Projects

DeepFake ECG introduces a GAN-based framework for generating realistic synthetic electrocardiogram signals. The project demonstrates that synthetic ECGs can augment training data for arrhythmia classifiers while preserving patient privacy, achieving performance on par with models trained on real clinical ECGs.

SEARCH pioneers healthcare innovation by combining synthetic data generation with federated learning to create a secure, privacy-preserving ecosystem for healthcare data sharing. The project addresses GDPR compliance and enables AI-driven clinical decision support across cardiovascular, gastrointestinal, and gynecological diseases, contributing to the European Health Data Space.

DADAP transforms psychiatric diagnostics for children and adolescents by leveraging AI, machine learning, and digitalization to enhance screening efficiency and throughput. The European consortium project aims to modernize diagnostic assessment workflows in child and adolescent psychiatry, improving access and reducing waiting times for mental health evaluations.

ECGDatasets is a community-maintained catalogue of 64+ publicly available ECG datasets from clinical hospitals, challenge competitions, and research institutions. It consolidates metadata scattered across PhysioNet, Zenodo, and Figshare into a single searchable reference, covering 12-lead, 2-lead, and 1-lead configurations. Includes Python analysis scripts and interactive D3.js visualisations for supported datasets.