URL study guide
https://tue.osiris-student.nl/onderwijscatalogus/extern/cursus?cursuscode=8BE130&collegejaar=2025&taal=enDescription
By 2025, artificial intelligence is a routine part of daily life, powering everything from autonomous vehicles to voice assistants and real-time information filtering. Meanwhile, healthcare systems face mounting pressure from a shrinking clinical workforce and an aging population. In response, hospitals and research institutions are accelerating the adoption of AI to support diagnostics, reduce workload and maintain consistent standards of care.Artificial intelligence has become central to modern medical image analysis, enabling more accurate, faster, and scalable interpretation of complex imaging data. Recent advances include the use of deep learning architectures such as vision transformers, as well as foundation models trained on large, diverse datasets. These systems achieve performance that approaches or exceeds expert-level accuracy in tasks such as disease diagnosing, segmentation, and prognostic prediction across multiple imaging modalities.
In this course, students will work with state-of-the-art AI methods, including language, vision, and multi-modal models, to address real-world medical image analysis problems. The focus is on practical applications that reflect current clinical needs, emphasizing both technical rigor and relevance to actual diagnostic workflows
Objectives
After passing the course, the student is able to:Understand the given medical image analysis problem and its clinical context;
Have insight in setting up a research question that can be quantitatively investigated;
Understand and apply AI-based image analysis algorithm(s) to facilitate this investigation;
Understand and motivate methodological choices;
Have insight in planning and documentation of the project progress;
Have insight in documentation and analysis of the solution and results of investigation.
Assessment (how the students will be assessed):
| Component | Insufficient | Satisfactory | Excellent |
| Methodology | Do not go much further than the minimal working example in the assignments, poor or incorrect methodological choices | Goes beyond the minimal working example in the assignment, well motivated and correct methodological choices | Use of state-of-the-art methodology |
| Experiments | Do not go much further than the minimal working example in the assignments, the experiments lack structure | Goes beyond the minimal working example in the assignment, the experiments are somewhat structured and systematic | Carefully considered set of experiments that demonstrate a systematic approach to the problem |
| Code | Missing or incomplete code structure, runs with errors, lacks documentation | Self-contained, does not result in errors, contains some documentation, can be easily used to reproduce the reported results | User-friendly, well-structured (good separation of general functionality and experiments), detailed documentation, optimized for speed, use of a version control system (such as GitHub) |