It would be interesting to know the gap between having 2D Xray as Input vs 3D CT as Input wrt to 3D Segmentation performance.
This will allow us to find how much performance is lost due to the projection from CT to X-ray.
Tasks
[ ] Define transformations for 3D Inputs, Use AttUnet as input
[ ] either use nnUnet or use the available pipeline
Description
It would be interesting to know the gap between having 2D Xray as Input vs 3D CT as Input wrt to 3D Segmentation performance. This will allow us to find how much performance is lost due to the projection from CT to X-ray.
Tasks