The lectures are divided into two mutually interconnected parts. In the first, theoretical part, students are introduced to the fundamental principles and methods of processing and analyzing image data used in biomedicine and clinical practice. Emphasis is placed on understanding mathematical foundations, algorithmic principles, and the logic of individual processing steps. In the second, application-oriented part, the theoretical knowledge is complemented by practical examples of procedures derived from real clinical practice and published scientific articles. Together with the instructor, students analyze specific methods, their limitations, approaches to evaluating results, and the appropriateness of their use for different types of medical data. This part also includes short student presentations based on selected published scientific articles. Their aim is to develop the ability to critically work with academic sources, understand modern methods, and navigate current trends in the field of medical image processing.

Lecture theory topics:

1. Introduction to biomedical imaging – overview of medical modalities and types of image data.
2. Discrete image and its representation – mathematical models, discretization, matrix representation.
3. Basic 2D transformations – Fourier transform, spatial and frequency filters.
4. Point and local operators – contrast adjustment, noise reduction, edge enhancement.
5. Feature extraction – textural features, convolution-based approaches.
6. Edge detection and edge representation – gradient operators, derivatives.
7. Image segmentation – thresholding, region-based methods, watershed, basics of machine learning.
8. Morphological operations – erosion, dilation, opening, closing.
9. Geometric transformations and image registration – rigid and nonlinear transformations.
10. Reconstruction from projections – backprojection, algebraic methods.

Application part of the lectures:

1. Practical examples from clinical practice (CT, MRI, ultrasound) – real tasks and procedures.
2. Analysis of published scientific articles – discussion of methods and results.
3. Advanced methods of modern research – e.g., deep learning, hybrid algorithms.
4. Student presentations of selected published articles.

PC exercises represent the practical component of the course and directly follow the topics covered in theoretical lectures. Each exercise begins with a short test verifying students’ understanding of the lecture content, thematically aligned with the topic of the given session. After the test, students work together with the instructor to implement and demonstrate practical applications of the discussed methods, solve typical problems in image processing, and examine how parameter settings influence the resulting outputs.

Students work individually on computers and complete tasks through practical programming in Python. The aim of the exercises is to develop the ability to apply image-processing methods in real tasks, understand the behavior of algorithms, and correctly interpret their outputs.

Outline of PC exercises:

1. Basic operations with discrete images, image spectrum; verification of Fourier transform properties.
2. Discrete 2D operators – point and local operations (contrast adjustment, gamma correction, edge enhancement, noise reduction).
3. Feature extraction – frequency-domain features, convolution-based features, texture analysis.
4. Edge detection – first- and second-derivative methods, combined approaches.
5. Image segmentation – simple, adaptive, multi- and semi-thresholding; Otsu’s method; region growing; watershed; parametric and geometric active contours; machine-learning approaches; morphological operations.
6. Object recognition – implementation of the Hough transform for line and circle detection.
7. 2D interpolation; implementation of rigid and affine geometric transformations; image registration and fusion.
8. Tomographic image reconstruction – inverse Radon transform approximation, sinogram construction, simple and filtered backprojection, fan-beam reconstruction.