Module CS3310-KP04

Duration

1 Semester

Turnus of offer

each winter semester

Credit points

4

Course of studies, specific fields and terms:

• Bachelor Robotics and Autonomous Systems 2016, optional subject, medical computer science

Classes and lectures:

• Medical Image Computing (exercise, 2 SWS)
• Medical Image Computing (lecture, 2 SWS)

Workload:

• 60 hours in-classroom work
• 40 hours private studies
• 20 hours exam preparation

Contents of teaching:

• Motivation, principles and applications of medical image computing
• Structure and formats of medical images
• Histograms and image transformations
• Image filtering using Fourier transform
• Image filtering with local operators
• Segmentation: thresholding, region growing
• Clusteranalysis and classifyer for image segmentation
• Introducing convolutional neural networks
• Morphological operators
• Application and evaluation of segmentation methods
• Image interpolation methods and transformataion of images
• Basic methods of image registration
• Combined signal and image analysis in 4D image processing

Qualification-goals/Competencies:

• Students are able to classify basic medical image processing methods, are able to characterize them and to apply them to concrete problems.
• They are able to select appropriate, problem-specific methods for image filtering, image segmentation, and morphological post-processing of segmentation results, to combine them in a processing pipeline and to use them for image enhancement or image segmentation of medical structures.
• They are able to distinguish between different methods of cluster analysis and statistical and neural pattern recognition and can characterize them based on different implicitly used model assumptions and properties.
• They are able to evaluate segmentation results of different methods based on established quality measures and to carry out an objective comparison of the quality of different segmentation methods in practical use.
• They are able to distinguish between different image interpolation methods, to classify them according to their specific advantages and disadvantages and to select an appropriate method and apply it, depending on a specific problem.
• They are able to assess the characteristics of different rigid image registration methods. For a specific registration problem they are able to select problem specific similarity measures and regularization terms and to parameterize them.
• They are able to distinguish and to characterize different techniques for analyzing functional 4D fMRI image sequences, with whom neurally activated brain areas in 4D image sequences of the head can be made visible.
• They are able to implement basic image processing algorithms and to bring them to use in combination with medical image processing modules available from program libraries.

Grading through:

• written exam

Is requisite for:

• Model and AI-based image processing in medicine (CS4332-KP06)
• Advanced Techniques of Medical Image Processing (CS4370-KP04, CS4370)
• Image Analysis and Visualization in Diagnostics and Therapy (CS4330-KP08, CS4330SJ14)

Requires:

• Analysis 2 (MA2500-KP04, MA2500)
• Analysis 1 (MA2000-KP08, MA2000)
• Linear Algebra and Discrete Structures 2 (MA1500-KP08, MA1500)
• Linear Algebra and Discrete Structures 1 (MA1000-KP08, MA1000)
• Introduction to Medical Informatics (CS1300-KP04, CS1300)

Responsible for this module:

• Prof. Dr. rer. nat. habil. Heinz Handels

Teacher:

• Institute of Medical Informatics
• Prof. Dr. rer. nat. habil. Heinz Handels
• Dr. rer. nat. Jan Ehrhardt

Literature:

• H. Handels : Medizinische Bildverarbeitung Stuttgart: Vieweg & Teubner 2009
• T. Lehmann : Handbuch der Medizinischen Informatik München: Hanser 2004
• M. Sonka, V. Hlavac, R. Boyle : Image Processing, Analysis and Machine Vision 2nd edition. Pacific Grove: PWS Publishing 1998

Language:

• offered only in German

Notes:

Last Updated:

11.03.2024