In this practical session, we will build a deep learning-based processing pipeline for cine cardiac magnetic resonance (CMR) image analysis and mesh fitting based on the ACDC dataset. This practical session will work through the following steps:
- Automated CMR image segmentation.
- CMR image-derived phenotypes calculation and analysis.
- Bi-ventricular heart mesh template construction.
- Bi-ventricular heart mesh fitting using a deep learning-based approach.
In Part I, we train a 2D U-Net model for cardiac MR segmentation. The practical notebook can be accessed in CMR_Segmentation_Demo.ipynb or via Google Colab link.

In Part II, we train a 3D U-Net model for cardiac mesh fitting. The practical notebook can be accessed in CMR_Mesh_Fit_Demo.ipynb or via Google Colab link.
