Imaging is a cornerstone of medicine. The number of medical imaging studies is growing rapidly, and so is the size and dimensionality of these images. Human experts interpret these images, which is time-consuming, expensive, and prone to errors because of visual fatigue. Advances in deep learning show that computers can extract more information from images, more reliably, and more accurately than ever before. However, most deep learning research in computer vision has focused on natural images. Adapting and further developing these techniques to the characteristics of medical images and medical data is an important and relevant research challenge.

Many conferences cover either medical imaging or machine learning. Many of them do cover the application of deep learning to medical imaging, often through satellite events, special session, and workshops. But there is currently no single venue which brings deep learning and medical imaging researchers together for in-depth discussion and exchange of ideas. With hundreds of deep learning papers being published in the field of medical imaging every year, and numerous AI-based startups in the medical field appearing, we believe such a venue is needed.

The MIDL conference aims to be a forum for deep learning researchers, clinicians and health-care companies to take a leap in the application of deep learning based automatic image analysis in disease screening, diagnosis, prognosis, treatment selection and treatment monitoring. The conference will have a broad scope and include topics such as computer-aided screening and diagnosis, detection, segmentation, (multi-modal) registration, image reconstruction and synthesis. Furthermore, we discuss issues such as the need for large curated and annotated datasets, noisy reference standards, and the high-dimensionality of medical data. Software demonstrations, presentation of medical data sets and innovative clinical applications are also covered as focus points for integration of deep learning algorithms in clinical practice.

MIDL currently offers a three-day program with keynote presentations from invited speakers, oral presentations, posters, and live demonstrations of deep learning algorithms from academia and industry.

We are committed to openness and transparency. We perform an open review process, have open access for all papers presented at MIDL, are transparent with regard to sponsorship packages and involvement from industry at the conference, provide freely available recordings of all presentations on the MIDL website, and we will urge MIDL contributors to use an open access policy as much as possible for the data and code.