Medical Image Analysis

The course covers computational techniques, from advanced mathematical modeling, machine and deep learning, and statistics, to design and validate image-guided medical procedures and computer-assisted diagnosis models. It highlights intrinsic and extrinsic information-based tracking, navigation, and the integration of pre- and intra-interventional data via non-rigid and 3D/2D image registration for procedure planning and visualization. The course delves into semantic image segmentation and quantitative analysis, exploring various methods including thresholding, clustering and mixture modeling, along with model- and atlas-based approaches. It introduces imaging biomarkers, with focus on their development and validation processes, and the field's latest advancements such as deep biomarker regression and geometrical learning on surface models of anatomy. All along the reference data acquisition and standardization procedures for clinical acceptance testing are covered.

The students will obtain a theoretical and practical insight into the computational and mathematical methods in medical image processing and analysis. Student will be able to design  automated methods for image enhancement and registration, segmentation and quantification of medical images, in the context of automatic image-based diagnosis and monitoring, and image-guided medical interventions.  They will have the capacity to design, acquire and organize the imaging data and reference data and measurements and use this data to objectively and critically assess the quality of image analysis methods.

Theoretical foundations of automated methods and exemplar use cases are given during lectures, while practical skills are gained through lab works, weekly labwork assignements and individual seminar work.