Philip Pentzel
Philip Pentzel
Thesis Topic: Application of the DCE-Movienet on the fastMRI-breast dataset and comparison to GRASP Reconstruction
Supervisors: Erik Gösche
Description
In the project for my bachelor’s thesis, the DCE-Movienet will be applied to the fastMRI-breast dataset. The aliased 4D-multi-coil images used as input of the Movienet will be computed from the breast dataset. The Movienet-CNN will be adjusted to fit the dimensions of the new input data and retrained. Reconstruction quality and speed will be compared to the GRASP reconstruction.
References
[1] Jing, J. et al. (2025). Combination of deep learning reconstruction and quantification for dynamic contrast-enhanced (DCE) MRI. (https://doi.org/10.1016/j.mri.2024.110310)
[2] Solomon, E. et al. (2025). FastMRI Breast: A Publicly Available Radial k-Space Dataset of Breast Dynamic Contrast-enhanced MRI. (https://doi/10.1148/ryai.240345).