Merge pull request #191 from mibaumgartner/nnunet_versio_bump

nnU-Net Version bump

nndet/inference/restore.py

@@ -106,7 +106,7 @@ def restore_fmap(fmap: np.ndarray,
         logger.info(f"Resampling: do separate z: {do_separate_z}; lowres axis: {lowres_axis}")
         fmap_old_spacing = resample_data_or_seg(fmap_transposed, size_after_cropping, is_seg=False,
                                                 axis=lowres_axis, order=interpolation_order,
-                                                do_separate_z=do_separate_z, cval=0,
+                                                do_separate_z=do_separate_z,
                                                 order_z=interpolation_order_z)
     else:
         logger.info(f"Resampling: no resampling necessary")

nndet/io/augmentation/bg_aug.py

@@ -155,10 +155,8 @@ class DefaultAug(NoAug):
             
             order_data=self.params.get("order_data"),
             border_mode_data=self.params.get("border_mode_data"),
-            border_cval_data=self.params.get("border_cval_data"),
             order_seg=self.params.get("order_seg"),
             border_mode_seg=self.params.get("border_mode_seg"),
-            border_cval_seg=self.params.get("border_cval_seg"),
             random_crop=self.params.get("random_crop"),
 
             p_el_per_sample=self.params.get("p_eldef"),
@@ -228,10 +226,8 @@ class BaseMoreAug(NoAug):
             
             order_data=self.params.get("order_data"),
             border_mode_data=self.params.get("border_mode_data"),
-            border_cval_data=self.params.get("border_cval_data"),
             order_seg=self.params.get("order_seg"),
             border_mode_seg=self.params.get("border_mode_seg"),
-            border_cval_seg=self.params.get("border_cval_seg"),
             random_crop=self.params.get("random_crop"),
 
             p_el_per_sample=self.params.get("p_eldef"),
@@ -323,10 +319,8 @@ class MoreAug(NoAug):
             
             order_data=self.params.get("order_data"),
             border_mode_data=self.params.get("border_mode_data"),
-            border_cval_data=self.params.get("border_cval_data"),
             order_seg=self.params.get("order_seg"),
             border_mode_seg=self.params.get("border_mode_seg"),
-            border_cval_seg=self.params.get("border_cval_seg"),
             random_crop=self.params.get("random_crop"),
 
             p_el_per_sample=self.params.get("p_eldef"),
@@ -430,10 +424,8 @@ class InsaneAug(NoAug):
             
             order_data=self.params.get("order_data"),
             border_mode_data=self.params.get("border_mode_data"),
-            border_cval_data=self.params.get("border_cval_data"),
             order_seg=self.params.get("order_seg"),
             border_mode_seg=self.params.get("border_mode_seg"),
-            border_cval_seg=self.params.get("border_cval_seg"),
             random_crop=self.params.get("random_crop"),
 
             p_el_per_sample=self.params.get("p_eldef"),

nndet/preprocessing/resampling.py

@@ -22,14 +22,14 @@ with SuppressPrint():
     import nnunet.preprocessing.preprocessing as nn_preprocessing
 
 
-def resize_segmentation(segmentation, new_shape, order=3, cval=0):
+def resize_segmentation(segmentation, new_shape, order=3):
     """
     Resizes a segmentation map. Supports all orders (see skimage documentation). Will transform segmentation map to one
     hot encoding which is resized and transformed back to a segmentation map.
     This prevents interpolation artifacts ([0, 0, 2] -> [0, 1, 2])
     """
     return nn_preprocessing.resize_segmentation(
-        segmentation=segmentation, new_shape=new_shape, order=order, cval=cval)
+        segmentation=segmentation, new_shape=new_shape, order=order)
 
 
 def get_do_separate_z(spacing, anisotropy_threshold: float = 3):
@@ -47,8 +47,6 @@ def resample_patient(data,
                      order_data=3,
                      order_seg=0,
                      force_separate_z=False,
-                     cval_data=0,
-                     cval_seg=-1,
                      order_z_data=0, 
                      order_z_seg=0,
                      separate_z_anisotropy_threshold: float = 3,
@@ -56,13 +54,13 @@ def resample_patient(data,
     return nn_preprocessing.resample_patient(data=data, seg=seg, original_spacing=original_spacing,
                                              target_spacing=target_spacing, order_data=order_data,
                                              order_seg=order_seg, force_separate_z=force_separate_z,
-                                             cval_data=cval_data, cval_seg=cval_seg, order_z_data=order_z_data,
+                                             order_z_data=order_z_data,
                                              order_z_seg=order_z_seg,
                                              separate_z_anisotropy_threshold=separate_z_anisotropy_threshold)
 
 
 def resample_data_or_seg(data, new_shape, is_seg, axis=None, order=3,
-                         do_separate_z=False, cval=0, order_z=0) -> np.ndarray:
+                         do_separate_z=False, order_z=0) -> np.ndarray:
     """
     Resample data or segmentation
 
@@ -73,7 +71,6 @@ def resample_data_or_seg(data, new_shape, is_seg, axis=None, order=3,
         axis: anisotropic axis, different resampling order used here
         order: order of resampling along the isotropic axis
         do_separate_z: Different resampling along z dimensions
-        cval: //
         order_z: if separate z resampling is done then this is the order for resampling in z
 
     Returns:
@@ -81,4 +78,4 @@ def resample_data_or_seg(data, new_shape, is_seg, axis=None, order=3,
     """
     return nn_preprocessing.resample_data_or_seg(
         data=data, new_shape=new_shape, is_seg=is_seg, axis=axis,
-        order=order, do_separate_z=do_separate_z, cval=cval, order_z=order_z)
+        order=order, do_separate_z=do_separate_z, order_z=order_z)

requirements.txt

 pytorch_lightning>=1.3.1,<=1.4.2
-batchgenerators==0.21
-nnunet==1.6.6
+batchgenerators>=0.24
+nnunet==1.7.1
 scipy
 scikit-learn
 scikit-image>=0.14