Optional check for more than one object in mask

extensions/chaeo/examples/transfer_labels_to_ilastik_object_classifier.py

@@ -33,6 +33,7 @@ def compare_object_maps(truth: GenericImageDataAccessor, inferred: GenericImageD
         if len(unique) == 1:  # no object in frame
             dd['inferred_label'] = unique[0]
         elif len(unique) > 2:  # multiple objects in frame, so mask out all but largest
+            print(f'Found multiple nonzero unique values in label {zi}: {unique}')
             ob_id = skimage.measure.label(inf_img)
             pr = skimage.measure.regionprops_table(ob_id, properties=['label', 'area'])
             mask = inf_img == pr['label'][pr['area'].argmax()]
@@ -43,6 +44,7 @@ def compare_object_maps(truth: GenericImageDataAccessor, inferred: GenericImageD
         labels.append(dd)
     return pd.DataFrame(labels)
 
+
 def infer_and_compare(classifier: PatchStackObjectClassifier, prefix, raw, mask, labels):
     result_acc, _ = classifier.infer(raw, mask)
     write_accessor_data_to_file(root / f'zstack_train_result.tif', result_acc)
@@ -72,24 +74,25 @@ if __name__ == '__main__':
         MonoPatchStackFromFile(root / 'zstack_train_mask.tif'),
         MonoPatchStackFromFile(root / 'zstack_train_label.tif'),
         label_names,
+        allow_multiple_objects=False
     )
     classifier = PatchStackObjectClassifier({'project_file': classifier_file})
 
-    # verify self-consistency of training set
-    infer_and_compare(
-        classifier,
-        'train',
-        MonoPatchStackFromFile(root / 'zstack_train_raw.tif'),
-        MonoPatchStackFromFile(root / 'zstack_train_mask.tif'),
-        MonoPatchStackFromFile(root / 'zstack_train_label.tif')
-    )
-
-    # run test set
-    infer_and_compare(
-        classifier,
-        'test',
-        MonoPatchStackFromFile(root / 'zstack_test_raw.tif'),
-        MonoPatchStackFromFile(root / 'zstack_test_mask.tif'),
-        MonoPatchStackFromFile(root / 'zstack_test_label.tif'),
-    )
+    # # verify self-consistency of training set
+    # infer_and_compare(
+    #     classifier,
+    #     'train',
+    #     MonoPatchStackFromFile(root / 'zstack_train_raw.tif'),
+    #     MonoPatchStackFromFile(root / 'zstack_train_mask.tif'),
+    #     MonoPatchStackFromFile(root / 'zstack_train_label.tif')
+    # )
+    #
+    # # run test set
+    # infer_and_compare(
+    #     classifier,
+    #     'test',
+    #     MonoPatchStackFromFile(root / 'zstack_test_raw.tif'),
+    #     MonoPatchStackFromFile(root / 'zstack_test_mask.tif'),
+    #     MonoPatchStackFromFile(root / 'zstack_test_label.tif'),
+    # )
 

extensions/chaeo/models.py

@@ -3,6 +3,7 @@ import shutil
 
 import h5py
 import numpy as np
+import skimage
 import vigra
 
 from extensions.chaeo.accessors import MonoPatchStack, MonoPatchStackFromFile
@@ -54,6 +55,7 @@ def generate_ilastik_object_classifier(
         label_stack: MonoPatchStackFromFile,
         label_names: list,
         lane: int = 0,
+        allow_multiple_objects=True,
 ) -> Path:
     """
     Starting with a template project file, transfer input data and labels to a new project file.
@@ -64,6 +66,7 @@ def generate_ilastik_object_classifier(
     :param label_stack: stack of patches containing object labels
     :param label_names: list of label names
     :param lane: ilastik lane identifier
+    :param allow_multiple_objects: skip check for multiple objects in each mask image
     :return: path to generated object classifier
     """
     assert mask_stack.shape == raw_stack.shape
@@ -85,6 +88,12 @@ def generate_ilastik_object_classifier(
         assert unique[-1] < len(label_names) + 1, f'Label ID {unique[-1]} exceeds number of label names: {len(label_names)}'
         labels.append(unique[-1])
 
+        if not allow_multiple_objects:
+            ob_id = skimage.measure.label(mask_stack.iat(ii))
+            num_obj = len(np.unique(ob_id)) - 1
+            if num_obj > 1:
+                raise MoreThanOneObjectError(f'Found {num_obj} connected objects in mask {ii}')
+
     # write to new project file
     with h5py.File(new_ilp, 'r+') as h5:
 
@@ -135,3 +144,8 @@ def generate_ilastik_object_classifier(
 
     return Path(new_ilp)
 
+class Error(Exception):
+    pass
+
+class MoreThanOneObjectError(Error):
+    pass
\ No newline at end of file