We are proud to announce the victory of our partner Transilvania University of Brasov (UTBV) as Category Winner of the EC’s Innovation Radar Prize – Industrial & Enabling Tech 2019, for the design and implementation of a software framework for developing privacy-preserving personalized medicine solutions based on homomorphically encrypted data and artificial intelligence (AI), particularly Deep Learning (DL), allowing full performance of AI models without compromising privacy.
The team is based at the Department of Automation and Information Technology, that works on the development on a wide perspective of healthcare-related applications, ranging from medical devices to medical applications (including physiological modelling, AI/DL-based image processing, high-performance computing, etc.).
About the innovation
Despite their potential in enabling personalized medicine applications, the adoption of DL-based solutions in clinical workflows has been hindered in many cases by strict regulations concerning the privacy of patient data. In the context of MHMD, the UTBV team has implemented a software framework for developing personalized medicine DL solutions based on homomorphically encrypted data. In a typical scenario, the client (hospital, patient, etc.) encrypts the data, sends it to the cloud-based application, which processes the data with the DL model and outputs encrypted results, which are then sent back to the client. Finally, the client decrypts and interprets the results. The model ensures that both input data and results remain private, and data analysis is performed only on the encrypted version of the data. Through this framework, UTBV has successfully y developed multiple personalized medicine applications, including: coronary artery disease assessment and diagnosis based on coronary angiographies (the hospital is the client), and long-term cardiovascular monitoring, relying on a reduced-order hemodynamic model and sensor data acquired through a wearable device (the patient is the client).