Training without patches #1179
Replies: 4 comments 3 replies
-
You are right, change the config/infer/train segmetnation_spleen scripts for modification. Specifically, change the for inference ROI, if your model is trained with different image size. |
Beta Was this translation helpful? Give feedback.
-
Thanks for your response. I want to confirm that "pixdim=(1.0, 1.0, 1.0)," means no interpolation..... |
Beta Was this translation helpful? Give feedback.
-
Thanks... want to confirm the units. 1.0 is definitely mm for all pixdim's? Not relative pixel width? |
Beta Was this translation helpful? Give feedback.
-
I think that I almost have it working. Any suggestions? def pre_transforms(self, data=None) -> Sequence[Callable]: def inferer(self, data=None) -> Inferer: def post_transforms(self, data=None) -> Sequence[Callable]: |
Beta Was this translation helpful? Give feedback.
-
I am using the Radiology App with segmentation_spleen to segment one structure from CT images.
I would like to train a 3D UNET without patches/ using all available data say 256x256x32.
If possible cropped from the center of the orig images.
I would appreciate any advice on what needs to be changed in the config/infer/train/etc files.
Beta Was this translation helpful? Give feedback.
All reactions