[WiSe 2024/25] Machine Learning in Medical Image Processing

Learning Outcomes

Participants will learn how to apply machine learning approaches such as support vector machines and deep learning for the automatic segmentation of medical image data.

Content

Clinical questions in image-based bloodflow analysis, requirement analysis based on a clinical application scenario, 3D image data preparation for machine learning, comparison and validation of image segmentation methods.

 

Description of Teaching and Learning Methods

In this project participants will work with CT angiography datasets typically acquired for the examination of bloodflow in the coronary artery. The goal is to implement and compare different machine learning approaches for the segmentation of these datasets. The sessions will consist of interleaved theoretical and practical tasks. The theoretical sessions will be dedicated to the introduction of the methods to apply for requirement analysis, data preprocessing, machine learning, validation and result presentation. In the practical sessions these methods will be applied to compile a concept for a clinical image analysis application, as well as an implementation and comparison of different machine learning approaches.

Exam information

Itermediate talks will take place during the semester and final talks will take place at the end of the winter semester 24/25.