Closed dev000000001 closed 4 months ago
In our 2dto3d module, we assign extrinsic camera parameters to their respective views according to the following scheme (x2ct_nerf/data/base.py
):
self.mapping_camera_type2pose = {
"PA": torch.tensor([0, 0]), # Posteroanterior view
"Lateral": torch.tensor([math.pi / 2, math.pi / 2]), # Lateral view
}
This configuration enables us to accurately replicate the camera's position and orientation for various types of DRR images, including posteroanterior (PA) and lateral views.
Hi,
Are you using the extrinsic camera parameters of the DRR image as input for perspective rendering?
Thanks