Research within KHeaT

The research group conducts research on deep learning methods for applications in the field of computer vision. Methodologically, two main research topics are currently in focus: 1) the investigation of machine learning methods with little or weakly annotated data (weakly supervised learning, semi-supervised learning, self-supervised learning). And 2) Exploring methods for modeling the uncertainty of learned (deep learning) models. In the field of health technologies we are working different problems related to medical image analysis, such as segmentation of MRT, PET, OCT, X-Ray images. Current research here addresses learning from sparsely annotated data, interactive data annotation, learning from images and text (medical reports). We are also interested in building assistive systems for sleep monitoring, or to detect stroke or epilepsia. An important part of our research addresses assistive technology to support blind and seeing impaired users.