Studienhandbuch | UE Deep Learning for Medical Imaging

Inhalt

Workload	Education level	Study areas	Responsible person	Hours per week	Coordinating university
1,5 ECTS	M2 - Master's programme 2. year	^(*)Artificial Intelligence	Erich Kobler	1 hpw	Johannes Kepler University Linz

Detailed information

Original study plan

Master's programme Artificial Intelligence 2025W

Learning Outcomes

Competences
Students gain an understanding of deep learning techniques applied to medical imaging. The course covers topics such as image generation for X-ray, CT, MRI, and ultrasound, and modern reconstruction methods. Students explore image segmentation, registration, solving inverse problems, and handling noisy data. They also develop skills to improve model robustness and understand federated learning for data privacy. Practical exercises showcase how deep learning impacts medical imaging in clinical practice.

Skills	Knowledge
Students know how to apply deep learning methods for image reconstruction, segmentation, and registration across medical imaging modalities. They can implement modern reconstruction techniques, solve inverse problems, and improve image quality. The course covers both classical and deep learning-based approaches. Students develop strategies to train models on noisy and sparse data while enhancing robustness against adversarial attacks and domain shifts. They also explore federated learning for secure, distributed model training.	By the end of the course, students understand medical imaging modalities and how deep learning can be integrated. They have learned about inverse problems and the evolution of reconstruction techniques. Additionally, students have insight into handling noisy data, improving robustness, and addressing domain adaptation. They also have explored federated learning and its role in secure model training.

Criteria for evaluation

Assignment with presentations and short paper

Methods

Group work on projects, slide presentations, discussion, report

Language

English

Changing subject?

On-site course
Maximum number of participants	-
Assignment procedure	Direct assignment