Signals and Systems Image processing and reconstruction

The goal of this course is to provide the students with a complete understanding of advanced techniques for image processing in the field of medical imaging. Using the concepts already learn in Image Processing, the student will be able to process images with techniques such as automatic segmentation, machine learning methods or pattern recognition. Image reconstruction from acquired data in CT, MR and Nuclear Medicine will also be covered. After completion of the course, the student should be able to select the proper approach to process medical image data depending on the modality and the desired output, to write the necessary program and to evaluate the results.

1. Review of basic concepts in image processing 2. Information Systems in the Medical Environment: DICOM, RIS and PACS. 3. 3D and 4D image visualization. Software tools for medical image analysis. 4. Wavelets and multi-resolution processing. 5. Advanced Segmentation I: detecting lines and shapes with Hough transform. 6. Image registration 7. Advanced Segmentation II:Adaptive filters 8. Advanced Segmentation III: Snakes and Active Contours 9. Tomographic Image Reconstruction in Projective Systems 10. Feature extraction and Machine Learning methods 11. Beyond classical methods: deep learning and convolutional networks.

Teaching methodology will be mainly based on lectures, seminars and practical sessions. Students are required to read assigned documentation before lectures and seminars. Lectures will be used by the teachers to stress and clarify some difficult or interesting points from the corresponding lesson, previously prepared by the student. Seminars will be mainly dedicated to interactive discussion with the students, present and evaluate homework. Grading will be based on continuous evaluation (including short-exams, homework, group essays, practical sessions, and student participation in class and Aula Global) and a final exam covering the whole subject. Help sessions and tutorial classes will be held prior to the final exam. Attendance to lectures, practical sessions, short-exams or submission of possible homework is not compulsory. However, failure to attend any exam or submit the exercises before the deadline will result in a mark of 0 in the corresponding continuous evaluation block. The practical sessions may consist on laboratory work or visits to research or clinical centers. A laboratory report will be required for each of them. Homework exercises will also be a very important contribution, since they will imply solving a specific problem, proposing an algorithm and implement it using computer tools. The attendance to 80% of practical sessions is mandatory. Failure to hand in the laboratory reports on time or unjustified lack of attendance will result in 0 marking for that continuous evaluation block. Some activities could reduce the total weight of the final exam, such as projects or open essays to be presented at the exam.