fix process ordering

ms_lesion_segmentation/animaMSLongitudinalPreprocessing.py

@@ -9,9 +9,9 @@
 
 # The preprocessing consists in three or four steps:
 #  - brain extraction
+#  - mask flair images with the union of the masks of both time points
 #  - bias correction
 #  - (optional) normalization on the given template
-#  - mask flair images with the union of the masks of both time points
 
 # Argument parsing
 parser = argparse.ArgumentParser(
@@ -83,9 +83,9 @@ def call(command):
 
 # Preprocess all patients: 
 #  - brain extraction
+#  - mask flair images with the union of the masks of both time points
 #  - bias correction
 #  - normalize (optional)
-#  - crop
 for patientName in os.listdir(patients):
 
     patient = os.path.join(patients, patientName)
@@ -101,8 +101,11 @@ def call(command):
 
     masks = []
 
-    # For both time points: extract brain and remove bias
-    for flairName in ['flair_time01_on_middle_space.nii.gz', 'flair_time02_on_middle_space.nii.gz']:
+    flairs = ['flair_time01_on_middle_space.nii.gz', 'flair_time02_on_middle_space.nii.gz']
+    groundTruths = ['ground_truth_expert1.nii.gz', 'ground_truth_expert2.nii.gz', 'ground_truth_expert3.nii.gz', 'ground_truth_expert4.nii.gz', 'ground_truth.nii.gz']
+
+    # For both time points: extract brain
+    for flairName in flairs:
 
         flair = os.path.join(patient, flairName)
         brain = os.path.join(patientOutput, flairName)
@@ -111,18 +114,8 @@ def call(command):
         # Extract brain
         call(["python", animaBrainExtraction, "-i", flair, "--mask", mask, "--brain", brain])
 
-        # Remove bias
-        call([animaN4BiasCorrection, "-i", brain, "-o", brain, "-B", "0.3"])
-
-        if templateFlair:
-            if os.path.exists(templateFlair):
-                # Normalize intensities with the given template
-                call([animaNyulStandardization, "-m", brain, "-r", templateFlair, "-o", brain])
-            else:
-                print('Template file ' + templateFlair + ' not found, skipping normalization.')
-
         masks.append(mask)
-    
+
     maskUnion = os.path.join(patientOutput, 'brain_mask.nii.gz')
 
     # Compute the union of the masks of both time points
@@ -132,15 +125,26 @@ def call(command):
     # Remove intermediate masks
     for mask in masks:
         os.remove(mask)
-
-    # Copy the ground truths to the output directory
-    for imageName in ['ground_truth_expert1.nii.gz', 'ground_truth_expert2.nii.gz', 'ground_truth_expert3.nii.gz', 'ground_truth_expert4.nii.gz', 'ground_truth.nii.gz']:
-        shutil.copyfile(os.path.join(patient, imageName), os.path.join(patientOutput, imageName))
 
-    # Mask original FLAIR images with the union mask
-    for flairName in ['flair_time01_on_middle_space.nii.gz', 'flair_time02_on_middle_space.nii.gz']:
+    # For both time points: mask, remove bias and normalize if necessary
+    for flairName in flairs:
 
         flair = os.path.join(patient, flairName)
         brain = os.path.join(patientOutput, flairName)
 
-        call([animaMaskImage, "-i", flair, "-m", maskUnion, "-o", brain])
+        # Mask original FLAIR images with the union mask
+        call([animaMaskImage, "-i", flair, "-m", maskUnion, "-o", brain])
+
+        # Remove bias
+        call([animaN4BiasCorrection, "-i", brain, "-o", brain, "-B", "0.3"])
+
+        if templateFlair:
+            if os.path.exists(templateFlair):
+                # Normalize intensities with the given template
+                call([animaNyulStandardization, "-m", brain, "-r", templateFlair, "-o", brain])
+            else:
+                print('Template file ' + templateFlair + ' not found, skipping normalization.')
+
+    # Copy the ground truths to the output directory
+    for imageName in groundTruths:
+        shutil.copyfile(os.path.join(patient, imageName), os.path.join(patientOutput, imageName))