Principles of Medical Image Analysis and AI

IIDS67541 Principles of Medical Image Analysis and AI 15 Semester 1 Postgraduate Taught Level 7 Arezoo Zakeri Unit coordinator Division of Informatics, Imaging and Data Sciences 1 FHEQ level (Framework for Higher Education Qualifications) ' Masters/Integrated Masters P4 ' European Credit Transfer & Accumulation System Rating : 7.5 Topics covered will include (but are not limited to):<ul><li>Basic image processing<ul><li>Image resampling, image filtering, noise models, region analysis  </li></ul></li><li>Methods for object detection and evaluation methods</li><li>Image segmentation methods and evaluation metrics</li><li>Image registration (linear and deformable approaches)</li><li>Deep learning (DL) for vision and representation learning<ul><li>Deep learning fundamentals  </li><li>Convolutional neural networks (CNN)</li><li>DL for medical image segmentation, object detection, and registration  </li></ul></li></ul> Topics covered will include (but are not limited to):<ul><li>Basic image processing<ul><li>Image resampling, image filtering, noise models, region analysis  </li></ul></li><li>Methods for object detection and evaluation methods</li><li>Image segmentation methods and evaluation metrics</li><li>Image registration (linear and deformable approaches)</li><li>Deep learning (DL) for vision and representation learning<ul><li>Deep learning fundamentals  </li><li>Convolutional neural networks (CNN)</li><li>DL for medical image segmentation, object detection, and registration  </li></ul></li></ul> The unit aims to:<ul><li>Introduce key image processing tools and methodologies for medical image analysis.</li><li>Introduce machine learning techniques for medical image interpretation and processing.</li><li>Introduce evaluation metrics for the assessment of medical image analysis methods  </li><li>Introduce key Python libraries for image analysis </li></ul> Students will be able to: <ul><li>apply fundamental digital image processing techniques that underpin medical image analysis, demonstrating an understanding of these methods and their role in medical imaging applications. Teaching will be delivered through lectures and practical sessions, with assessment via a summative assignment and a final examination.</li><li>apply core deep learning principles to medical image computing tasks such as image segmentation and object detection, demonstrating an understanding of these methods and how they are used in medical imaging applications. Teaching will be delivered through lectures and practical sessions, with assessment via a summative assignment and a final examination.  </li><li>evaluate medical image analysis methods using appropriate performance metrics and interpret the results in scientific reports. Assessment will be via the summative assignment. </li></ul> Students will be able to: <ul><li>select and apply appropriate techniques for medical image processing, including intensity processing; frequency processing, spatial and spectral filtering; morphological processing. Assessment will be via the summative assignment. </li></ul> Students will be able to: <ul><li>modify Python code provided for practical image processing tasks using different deep learning models.</li><li>critically assess the impact of the changes on the performance based upon knowledge of intended outcome and evaluation methodology. Assessment will be via the summative assignment. </li></ul> Students will be able to: <ul><li>undertake a project in medical image analysis  </li><li>write a scientific report. Assessment will be via the summative assignment. </li></ul> The course will be delivered as either in person 12 x 90 minute lectures, each typically supplemented with 15-30 minutes of discussion and followed by a 60-minute practical session.Lectures will be reinforced by practical exercises backed up with online reading material in the form of research papers.  Exercises will be given at the end of each lecture in order to introduce intellectual content via task-based learning. Solutions will be discussed prior to delivery of each week’s new material, so that students receive regular feedback.  Towards the end of the semester, the understanding of medical image processing techniques will be assessed in a written summative assignment. The summative assignment on AI-based models examines how modifications—such as image pre-processing, data preparation, network architecture, optimizer selection, and parameter tuning—affect model performance and efficiency, using evaluation metrics. 1 Written exam 65% 2 Written assignment (inc essay) 35% <figure class="table"><table><thead><tr><th>Assessment task</th><th>Length</th><th>Weighting within unit</th></tr></thead><tbody><tr><td>Written summative assignment</td><td>1500 words</td><td>35%</td></tr><tr><td>Final exam</td><td>1 hour</td><td>65%</td></tr></tbody></table></figure> Written feedback  and/or marks will be provided within the 15 working day policy specified at the University. MSc Applied AI for Med Imaging MSc Applied AI for Med Imaging Not Set Mandatory N The material for this course comes from a variety of sources and is not available in a single reference text. However, some parts of the material are covered in the following:<ul><li>Medical Image Analysis ed. Frangi et al. 2024.</li><li>Guide to Medical Image Analysis_ Methods and Algorithms_ Advances in Computer Vision and Pattern Recognition [Toennies 2012-02-06]</li><li>Handbook of Medical Imaging, Processing and Analysis, Isaac N. Bankman (Editor in chief), Academic Press, 2000.</li></ul>Further reading:<ul><li>J. Kaur and W. Singh, “Tools, techniques, datasets and application areas for object detection in an image: a review,” Multimedia Tools and Applications, 2022. DOI: 10.1007/s11042-022-13153-y</li><li>F. P. M. Oliveira and J. M. R. S. Tavares, “Medical image registration: A review,” Computer Methods In Biomechanics & Bio Engineering, 2014, 17(2):73-93. DOI: 10.1080/10255842.2012.670855</li><li>A S Lundervold and A. Lundervold “An overview of deep learning in medical imaging focusing on MRI,” Zeitschrift für Medizinische Physik, 2018.https://doi.org/10.1016/j.zemedi.2018.11.002 </li></ul> Practical classes & workshops 30 Seminars 36 Tutorials 24 60