Medical Imaging with AI Integration (ELE670)

Medical data, in the form of images, is largely used as an important part of the diagnostics. This course deals with some key techniques for collecting such data. The theme is seen in relation to image processing as well as machine learning/ deep learning , which are core subjects in the study program, as such methods can be used for automatic segmentation, interpretation and analysis of images. In modern diagnostics, automatic data analysis can be included as decision support.

The the following imaging techniques will be emphasized: , X-ray, Magnetic Resonance Imaging (MR), Computer Tomography (CT), Angiography, Ultrasound. Also it will cover using AI in medical image analysis, medical image acquisition using deep learning, dealing with medical data imbalance and modern neural networks in the field of medical diagnostic.

Note! This is an elective course and may be cancelled if fewer than 10 students are enrolled by August 20 for the autumn semester/January 20 for the spring semester. 

This course addresses how some selected medical images are formed and the characteristics of these. This is, to some extent, seen in context with the themes and techniques of image processing and machine learning and artificial intelligence.

The course will focus on principles, modes of operation, applications, and study of example images for some common techniques for collecting medical diagnostic data. The following techniques will be highlighted:

- X-ray, mammography, angiography

- Magnetic resonance imaging (MR)

- Computer tomography (CT)

- Pathology and microscopic imaging

- Basic ideas of pixel/voxel, metadata, some tours of visualisation tools

- Standard formats (DICOM vs NIfTI) and large public dataset banks.

- Integration of AI in medical images, analysis and acquisition

- Modern medical image segmentation networks (U-Net and transformer-based models)

- Medical image reconstruction

- Multiple instance learning

- Handling imbalanced data

Knowledge:

The purpose of the course is to provide students with a technological background insight into techniques for the formation of medical diagnostic images and integration of AI applied to the data. Such medical data should then be seen in the context of techniques and knowledge from other subjects. Students will learn about a number of different techniques for collecting medical diagnostic data. The following will be emphasized: x-ray, magnetic resonance imaging (MRI), computer tomography (CT), ultrasound, angiography, microscopy etc and how to use AI for better data collection and analysis. Students will learn about the principles, operations and applications of these techniques, for example by means of sample images.

Skills:

The students should be able to explain the principles behind some techniques for collecting medical diagnostic signals and images. The student should be able to recognize and understand the meaning of specific characteristics from different types of images and how artificial intelligence is used to analyze them and what are its effects. Prepare medical data to be used as an input to a neural network and set up/choose suitable networks based on their applications.

General competence:

After taking this course, students will be able to understand the connection between medical diagnostic images and physiological phenomena.

To take this course, you are expected to have a bachelor’s degree in engineering or technology (e.g., biomedical engineering, electrical engineering, or computer science). It is expected that you have had a course in programming as Python programming must be used in mandatory assignments and project work. Mathematical knowledge at the level of a completed engineering degree is required, including linear algebra with matrices and vector calculus.

Familiarity with concepts from deep learning such as neural networks, loss functions, and model training is recommended, though the course includes a brief introduction to these topics.

Prior exposure to medical imaging is helpful but not required, as the fundamentals of radiography, CT, MRI, and ultrasound will be covered in the first part of the course.