AI for Health Imaging

The project will implement a cloud-based platform, based on Open Source and open standards, providing, via automated APIs, secure and anonymous/pseudo-anonymous service to the research and medical community all over Europe for automated analysis of images potentially indicating skin cancer disease. Advanced techniques for anonymization, information security and protection based on blockchain, as well as ethical and GDPR compliance will be included by design. Within the project, algorithms based on artificial intelligence will be developed and validated for their use on the interpretation of pathological images for the diagnosis of skin cancer. A significant requirement for the success of such algorithms and the subsequent use of the platform infrastructure is the population of the database and testing of the algorithms through large sets of data. The commitment of end-users, a network of European hospitals, with this project will ensure the provision of a large database of pathological images for the development, validation and final implementation of the AI database and platform. The involvement of end-users throughout the project will at the same time ensure the relevance of the analysis to the day-to-day jobs of pathologists studying skin cancer, not only on the information extracted from the images but also on the functioning, requirements, user-experience and complexity of the final system. The implementation of the platform in return will present the database as a reference in Europe, allowing for clinicians and researchers to utilise the validated artificial intelligence algorithms for improved and faster identification, diagnosis and treatment of cancer