- Voxel-based image segmentation
- Graph-based image segmentation
- Contour-based image segmentation
- Model-based image segmentation
- Image segmentation with deep learning
The course consists of lectures, laboratories, mathematical exercises, and an exam. Participants combine basic and advanced software libraries for image registration in Python, including scipy, numpy, SimpleITK, scikit-image, etc. Some specific labs use MATLAB and Mialab, an image segmentation tool developed at KTH. The course also includes introductory labs for students with programming experience but no Python experience.
Image segmentation is used to identify relevant regions in images. Image segmentation is important for the diagnosis and treatment of various diseases. The course covers concepts, theories, and the most used methods in image segmentation. The course is focused on solving medically relevant problems.
After completing the course, the participant should be able to:
- Understand the key issues and challenges in image segmentation
- Describe the main principles and methods and the main differences between them
- Summarize the advantages and disadvantages and scope of different methods
- Identify and understand the mathematical theory behind the most used methods
- Develop and systematically evaluate different methods for solving simplified problems
- Analyze the effect of different parameters of the methods in particular situations
- Explain the proposed strategy for solving specific problems
in order to:
- understand the complete workflow for using computational tools for image segmentation in a medical context
- be able to implement computational solutions in image segmentation for medically relevant problems
- have a broad knowledge base that can facilitate understanding literature in the field
