Merge pull request #134 from loli/Release_0.5.2

Release 0.5.2

CHANGES.txt

@@ -1,3 +1,4 @@
+v0.5.2, 2024-07-23 -- Hotfixes
 v0.5.1, 2024-04-03 -- Hotfixes
 v0.5.0, 2024-04-03 -- Addressed all depreciation warnings and incompatabilities
                       Updated documentation

doc/source/conf.py

@@ -50,7 +50,7 @@
 # The short X.Y version.
 version = "0.5"
 # The full version, including alpha/beta/rc tags.
-release = "0.5.1"
+release = "0.5.2"
 
 # Automatically created autosummary entries (thus no need to call sphinx-autogen)
 autosummary_generate = True

medpy/__init__.py

@@ -24,4 +24,4 @@
 along with this program.  If not, see <http://www.gnu.org/licenses/>.
 """
 
-__version__ = "0.5.1"
+__version__ = "0.5.2"

medpy/features/intensity.py

@@ -218,7 +218,7 @@ def indices(image, voxelspacing=None, mask=slice(None)):
         image = image[0]
 
     if not type(mask) is slice:
-        mask = numpy.array(mask, copy=False, dtype=numpy.bool_)
+        numpy.asarray(mask).astype(bool)
 
     if voxelspacing is None:
         voxelspacing = [1.0] * image.ndim
@@ -865,7 +865,7 @@ def _extract_centerdistance(image, mask=slice(None), voxelspacing=None):
     """
     Internal, single-image version of `centerdistance`.
     """
-    image = numpy.array(image, copy=False)
+    image = numpy.asarray(image)
 
     if None == voxelspacing:
         voxelspacing = [1.0] * image.ndim
@@ -889,7 +889,7 @@ def _extract_intensities(image, mask=slice(None)):
     """
     if type(mask) is list and type(mask[0]) is slice:
         mask = tuple(mask)
-    return numpy.array(image, copy=True)[mask].ravel()
+    return numpy.asarray(image)[mask].ravel()
 
 
 def _substract_hemispheres(
@@ -941,7 +941,7 @@ def _extract_feature(fun, image, mask=slice(None), **kwargs):
         Additional keyword arguments to be passed to the feature extraction function
     """
     if not type(mask) is slice:
-        mask = numpy.array(mask, copy=False, dtype=numpy.bool_)
+        numpy.asarray(mask).astype(bool)
 
     if type(image) is tuple or type(image) is list:
         return join(*[fun(i, mask, **kwargs) for i in image])

medpy/features/texture.py

@@ -78,7 +78,7 @@ def coarseness(image, voxelspacing=None, mask=slice(None)):
     # set default mask or apply given mask
     if not type(mask) is slice:
         if not type(mask[0] is slice):
-            mask = numpy.array(mask, copy=False, dtype=numpy.bool_)
+            mask = numpy.asarray(mask).astype(bool)
     image = image[mask]
 
     # set default voxel spacing if not suppliec
@@ -170,7 +170,7 @@ def contrast(image, mask=slice(None)):
     # set default mask or apply given mask
     if not type(mask) is slice:
         if not type(mask[0] is slice):
-            mask = numpy.array(mask, copy=False, dtype=numpy.bool_)
+            mask = numpy.asarray(mask).astype(bool)
         else:
             mask = tuple(mask)
     image = image[mask]
@@ -238,7 +238,7 @@ def directionality(
     # set default mask or apply given mask
     if not type(mask) is slice:
         if not type(mask[0] is slice):
-            mask = numpy.array(mask, copy=False, dtype=numpy.bool_)
+            mask = numpy.asarray(mask).astype(bool)
     image = image[mask]
 
     # set default voxel spacing if not suppliec

medpy/features/utilities.py

@@ -77,7 +77,7 @@ def normalize(vector, cutoffp=(0, 100), model=False):
         The learned normalization model.
 
     """
-    vector = numpy.array(vector, dtype=float)
+    vector = numpy.asarray(vector).astype(float)
 
     # add a singleton dimension if required
     if 1 == vector.ndim:
@@ -121,7 +121,7 @@ def normalize_with_model(vector, model):
     normalize : ndarray
         The normalized versions of the input vectors.
     """
-    vector = numpy.array(vector, dtype=float)
+    vector = numpy.asarray(vector).astype(float)
 
     # unpack model
     minp, maxp, minv, maxv = model
@@ -213,9 +213,9 @@ def join(*vectors):
     # process supplied arguments
     vectors = list(vectors)
     for i in range(len(vectors)):
-        vectors[i] = numpy.array(vectors[i], copy=False)
+        vectors[i] = numpy.asarray(vectors[i])
         if vectors[i].ndim == 1:
-            vectors[i] = numpy.array([vectors[i]], copy=False).T
+            vectors[i] = numpy.asarray([vectors[i]]).T
 
     # treat single-value cases special (no squeezing)
     if 1 == len(vectors[0]):

medpy/metric/binary.py

@@ -467,12 +467,12 @@ def assd(result, reference, voxelspacing=None, connectivity=1):
 
     and then averaging the two lists. The binary images can therefore be supplied in any order.
     """
-    assd = numpy.mean(
-        (
+    assd = numpy.concatenate(
+        [
             __surface_distances(result, reference, voxelspacing, connectivity),
             __surface_distances(reference, result, voxelspacing, connectivity),
-        )
-    )
+        ]
+    ).mean()
     return assd
 
 
@@ -802,12 +802,12 @@ def obj_assd(result, reference, voxelspacing=None, connectivity=1):
 
     and then averaging the two lists. The binary images can therefore be supplied in any order.
     """
-    assd = numpy.mean(
-        (
+    assd = numpy.concatenate(
+        [
             __obj_surface_distances(result, reference, voxelspacing, connectivity),
             __obj_surface_distances(reference, result, voxelspacing, connectivity),
-        )
-    )
+        ]
+    ).mean()
     return assd
 
 

setup.py

@@ -138,7 +138,7 @@ def run_setup(with_compilation):
 
     setup(
         name="MedPy",
-        version="0.5.1",  # major.minor.micro
+        version="0.5.2",  # major.minor.micro
         description="Medical image processing in Python",
         author="Oskar Maier",
         author_email="oskar.maier@gmail.com",

tests/metric_/binary.py

@@ -0,0 +1,77 @@
+"""
+Unittest for medpy.features.histogram.
+
+@author Oskar Maier
+@version r0.1.0
+@since 2024-07-23
+@status Release
+"""
+
+import numpy as np
+
+from medpy.metric import asd, assd, obj_asd, obj_assd
+
+result_min = np.asarray([1, 0]).astype(bool)
+reference_min = np.asarray([0, 1]).astype(bool)
+result_sym = np.asarray(
+    [[1, 0, 0, 0], [1, 0, 0, 0], [1, 0, 0, 0], [1, 1, 0, 0]]
+).astype(bool)
+reference_sym = np.asarray(
+    [[1, 1, 0, 0], [1, 0, 0, 0], [1, 0, 0, 0], [0, 0, 0, 0]]
+).astype(bool)
+
+
+def test_asd_identity():
+    assert asd(result_min, result_min) == 0
+
+
+def test_assd_identity():
+    assert assd(result_min, result_min) == 0
+
+
+def test_obj_asd_identity():
+    assert obj_asd(result_min, result_min) == 0
+
+
+def test_obj_assd_identity():
+    assert obj_assd(result_min, result_min) == 0
+
+
+def test_asd_distance():
+    assert asd(result_min, reference_min) == 1.0
+
+
+def test_assd_distance():
+    assert assd(result_min, reference_min) == 1.0
+
+
+def test_asd_voxelspacing():
+    assert asd(result_min, reference_min, voxelspacing=[2]) == 2.0
+
+
+def test_assd_voxelspacing():
+    assert assd(result_min, reference_min, voxelspacing=[2]) == 2.0
+
+
+def test_asd_is_not_symetric():
+    asd_1 = asd(result_sym, reference_sym)
+    asd_2 = asd(reference_sym, result_sym)
+    assert asd_1 != asd_2
+
+
+def test_assd_is_symetric():
+    assd_1 = assd(result_sym, reference_sym)
+    assd_2 = assd(reference_sym, result_sym)
+    assert assd_1 == assd_2
+
+
+def test_obj_asd_is_not_symetric():
+    asd_1 = obj_asd(result_sym, reference_sym)
+    asd_2 = obj_asd(reference_sym, result_sym)
+    assert asd_1 != asd_2
+
+
+def test_obj_assd_is_symetric():
+    assd_1 = obj_assd(result_sym, reference_sym)
+    assd_2 = obj_assd(reference_sym, result_sym)
+    assert assd_1 == assd_2