Merge pull request #73 from tonysyu/fix-ctmf

BUG: Force median filter parameters to be valid.

doc/source/api_changes.txt

@@ -3,3 +3,7 @@ Version 0.3
 - Remove ``as_grey``, ``dtype`` keyword from ImageCollection
 - Remove ``dtype`` from imread
 - Generalise ImageCollection to accept a load_func
+
+Version 0.4
+-----------
+- Switch mask and radius arguments for median_filter

skimage/filter/ctmf.py

@@ -16,20 +16,20 @@ from . import _ctmf
 from rank_order import rank_order
 
 
-def median_filter(image, mask=None, radius=2, percent=50):
+def median_filter(image, radius=2, mask=None, percent=50):
     '''Masked median filter with octagon shape.
 
     Parameters
     ----------
     image : (M,N) ndarray, dtype uint8
         Input image.
+    radius : {int, 1}, optional
+        The radius of a circle inscribed into the filtering
+        octagon. Default radius is 1.
     mask : (M,N) ndarray, dtype uint8, optional
         A value of 1 indicates a significant pixel, 0
         that a pixel is masked.  By default, all pixels
         are considered.
-    radius : {int, 1}, optional
-        The radius of a circle inscribed into the filtering
-        octagon. Default radius is 1.
     percent : {int, 50}, optional
         The unmasked pixels within the octagon are sorted, and the
         value at the `percent`-th index chosen.  For example, the
@@ -47,6 +47,9 @@ def median_filter(image, mask=None, radius=2, percent=50):
     if image.ndim != 2:
         raise TypeError("The input 'image' must be a two dimensional array.")
 
+    if radius < 2:
+        raise ValueError("The input 'radius' must be >= 2.")
+
     if mask is None:
         mask = np.ones(image.shape, dtype=np.bool)
     mask = np.ascontiguousarray(mask, dtype=np.bool)

skimage/filter/tests/test_ctmf.py

@@ -1,11 +1,12 @@
 import numpy as np
+from nose.tools import raises
 
 from skimage.filter import median_filter
 
 
 def test_00_00_zeros():
     '''The median filter on an array of all zeros should be zero'''
-    result = median_filter(np.zeros((10, 10)), np.ones((10, 10), bool), 3)
+    result = median_filter(np.zeros((10, 10)), 3, np.ones((10, 10), bool))
     assert np.all(result == 0)
 
 
@@ -13,14 +14,14 @@ def test_00_01_all_masked():
     '''Test a completely masked image
 
     Regression test of IMG-1029'''
-    result = median_filter(np.zeros((10, 10)), np.zeros((10, 10), bool), 3)
+    result = median_filter(np.zeros((10, 10)), 3, np.zeros((10, 10), bool))
     assert (np.all(result == 0))
 
 
 def test_00_02_all_but_one_masked():
     mask = np.zeros((10, 10), bool)
     mask[5, 5] = True
-    median_filter(np.zeros((10, 10)), mask, 3)
+    median_filter(np.zeros((10, 10)), 3, mask)
 
 
 def test_01_01_mask():
@@ -29,7 +30,7 @@ def test_01_01_mask():
     img[5, 5] = 1
     mask = np.ones((10, 10), bool)
     mask[5, 5] = False
-    result = median_filter(img, mask, 3)
+    result = median_filter(img, 3, mask)
     assert (np.all(result[mask] == 0))
     np.testing.assert_equal(result[5, 5], 1)
 
@@ -38,7 +39,7 @@ def test_02_01_median():
     '''A median filter larger than the image = median of image'''
     np.random.seed(0)
     img = np.random.uniform(size=(9, 9))
-    result = median_filter(img, np.ones((9, 9), bool), 20)
+    result = median_filter(img, 20, np.ones((9, 9), bool))
     np.testing.assert_equal(result[0, 0], np.median(img))
     assert (np.all(result == np.median(img)))
 
@@ -47,7 +48,7 @@ def test_02_02_median_bigger():
     '''Use an image of more than 255 values to test approximation'''
     np.random.seed(0)
     img = np.random.uniform(size=(20, 20))
-    result = median_filter(img, np.ones((20, 20), bool), 40)
+    result = median_filter(img, 40, np.ones((20, 20), bool))
     sorted = np.ravel(img)
     sorted.sort()
     min_acceptable = sorted[198]
@@ -77,7 +78,7 @@ def test_03_01_shape():
     octagon[i - j > radius + a_2] = False
     np.random.seed(0)
     img = np.random.uniform(size=(21, 21))
-    result = median_filter(img, np.ones((21, 21), bool), radius)
+    result = median_filter(img, radius, np.ones((21, 21), bool))
     sorted = img[octagon]
     sorted.sort()
     min_acceptable = sorted[len(sorted) / 2 - 1]
@@ -93,7 +94,7 @@ def test_04_01_half_masked():
     mask[10:, :] = False
     img[~ mask] = 2
     img[1, 1] = 0  # to prevent short circuit for uniform data.
-    result = median_filter(img, mask, 5)
+    result = median_filter(img, 5, mask)
     # in partial coverage areas, the result should be only
     # from the masked pixels
     assert (np.all(result[:14, :] == 1))
@@ -105,10 +106,16 @@ def test_04_01_half_masked():
 def test_default_values():
     img = (np.random.random((20, 20)) * 255).astype(np.uint8)
     mask = np.ones((20, 20), dtype=np.uint8)
-    result1 = median_filter(img, mask, radius=2, percent=50)
+    result1 = median_filter(img, radius=2, mask=mask, percent=50)
     result2 = median_filter(img)
     np.testing.assert_array_equal(result1, result2)
 
 
+@raises(ValueError)
+def test_insufficient_size():
+    img = (np.random.random((20, 20)) * 255).astype(np.uint8)
+    median_filter(img, radius=1)
+
+
 if __name__ == "__main__":
     np.testing.run_module_suite()