AI Development

This WG focuses on image quantification, distributed learning, and interpretability within the field of AI. It plays a crucial role in defining a strategic position on AI development and engages in discussions around common questions. Furthermore, the WG organizes its activities around a central topic aligned with the projects being addressed. This allows for comprehensive discussions and comparisons of potential solutions in areas such as data harmonization, domain adaptation strategies, unlabelled data exploitation, and uncertainty estimation. * The WG has developed a questionnaire to collect experiences and comments from AI developers. This questionnaire contributes to the understanding of AI development and assessing the methods, tools and software are used in each step of the radiomics workflow. Additionally, it seeks to determine whether these methods, tools, or software are open source, fostering transparency within the field. This WG has also been working on the preparation of a manuscript that aims to explore and put together the experience on medical image annotation from the EU research projects which are part of the AI for Health Imaging (AI4HI) network (CHAIMELEON, EuCanImage, INCISIVE, ProCAncer-I, PRIMAGE). The final aim is to propose ontologies, standards, tools and recommendations to help the medical imaging community to perform automatic high-quality annotations in medical images.