Teaching medical image analysis requires innovative approaches to enhance student engagement and align learning outcomes with real-world challenges. This paper presents the implementation of active learning strategies across multiple modules in a medical image analysis curriculum: Medical Image Registration and Applications (MIRA), Computer-Aided Diagnosis (CAD), Medical Image Segmentation and Applications (MISA), and E-health. These interconnected modules progressively build student knowledge through problem-based learning (PBL), collaborative labs, flipped classrooms, and peer learning.

In MIRA, PBL is used to engage students in a real research challenge—image registration of chest CT images. Throughout the course, students work on this project, producing results that align with state-of-the-art research, fostering critical thinking, teamwork, and research-oriented learning. MISA follows a similar model, where students implement segmentation algorithms for medical imaging.

To enhance interdisciplinary learning, we introduce a collaborative lab where students apply non-rigid registration (MIRA) and Expectation-Maximization segmentation (MISA) in atlas-based segmentation. This activity bridges theoretical concepts with practical implementation, fostering hands-on learning and interdisciplinary engagement.

The CAD module explores AI and image processing techniques for medical diagnosis, where students develop computer-aided tools. The course covers feature extraction, machine learning, and deep learning for disease classification. Some deep learning concepts are further explored in MISA, particularly in 3D brain MRI segmentation. Both modules emphasize performance evaluation using metrics like sensitivity, specificity, AUC-ROC, Dice, and Hausdorff distance, reinforcing the importance of quantitative assessment in lab sessions.

In the E-health module, flipped classrooms and peer learning are employed. The course is divided into two parts: first, medical professionals (radiologists, neurologists, cardiologists, neuropsychologists) present real-world applications of medical imaging, and students visit radiology departments. In the second part, students work in teams to design and implement lab sessions for their peers, selecting a topic with a medical imaging component and creating hands-on exercises. This structure encourages autonomy, feedback-driven learning, and collaboration.

The innovative methodologies presented in this paper significantly enhance student engagement, comprehension, and performance in medical image analysis. By integrating research-driven challenges, collaborative experiences, and student-led activities, we provide a dynamic and effective educational framework. Quantitative assessments reveal high student satisfaction, reflected in course evaluations and surveys. Students consistently rate the modules with top scores, emphasizing the effectiveness of active learning in improving their understanding and practical skills. Additionally, employability outcomes are strong, with all graduates securing positions in academia (PhDs) or industry shortly after completing the master’s program.

Keywords: Technology, Active learning, medical image analysis, Problem based learning, peer learning.