Semi-supervised Learning for Medical Image Analysis

Research Project, 2020 – 2025

Most recent successes of machine learning have been based on Supervised Learning (SL) methods, fueled by large quantities of parallel compute power and humanly annotated training data. However, that option quickly becomes intractable due to the labour intensive work of manual annotation, especially for medical image data. Instead, many believe that Semi-Supervised Learning (SSL) will drive the next AI revolution by using vast amount of unlabeled data (and some labeled examples) to discover all concepts and underlying causes that matter when interpreting an image. In this project, we will develop new methods and techniques for SSL and apply it to medically relevant problems where lots of image data is available. 

Two medical image domains have been identified, but we expect that our techniques will be applicable to other domains as well. The first one is based on the SCAPIS study, a population study with collected CT examinations of over 28,000 individuals. These examinations can be used to analyze, for instance, possible atherosclerosis in the coronary arteries, which in turn can predict the risk of myocardial infarction in the future. The second concerns automatic analysis and diagnosis of cardiac ultrasound images. Currently, there are 90.000 of so called echocardiographies collected. The research will be performed in close collaboration with medical researchers from Sahlgrenska Academy at Gothenburg University.