Medical Imaging

Image C.V.

Under development.

Course objective

Course content

• PET basics (more than SPECT) up to classical iterative reconstruction, including uncertainty estimation;
• Tomographic reconstruction methods based on machine learning, in particular Plug&Play and Unrolling.
• Lab session :
  • classical reconstruction (iterative)
  • Plug&Play reconstruction in PET
• MRI concepts and basic principles of MRI data acquisition and image reconstruction
• Classical approaches to accelerated MRI imaging: parallel imaging, compressed sensing, non-Cartesian imaging
• Supervised and unsupervised learning for MRI image reconstruction, Non-Cartesian trajectory learning.
• Lab session
  • Effects of Cartesian and non-Cartesian subsampling in accelerated imaging
  • Image reconstruction in parallel imaging and compressed sensing
• Lab session
  • Deep learning for MRI image reconstruction: unrolled vs. PnP approaches

References