FIX add missing min_size parameter, add regression test.

skimage/segmentation/_felzenszwalb.py

@@ -43,10 +43,9 @@ def felzenszwalb(image, scale=1, sigma=0.8, min_size=20):
            Huttenlocher, D.P.  International Journal of Computer Vision, 2004
     """
 
-    #image = img_as_float(image)
     if image.ndim == 2:
         # assume single channel image
-        return _felzenszwalb_grey(image, scale=scale, sigma=sigma)
+        return _felzenszwalb_grey(image, scale=scale, sigma=sigma, min_size=min_size)
 
     elif image.ndim != 3:
         raise ValueError("Felzenswalb segmentation can only operate on RGB and"

skimage/segmentation/tests/test_felzenszwalb.py

@@ -1,7 +1,9 @@
 import numpy as np
 from numpy.testing import assert_equal, assert_array_equal
+
 from skimage._shared.testing import assert_greater
 from skimage.segmentation import felzenszwalb
+from skimage import data
 
 
 def test_grey():
@@ -18,6 +20,24 @@ def test_grey():
         hist = np.histogram(img[seg == i], bins=[0, 0.1, 0.3, 0.5, 1])[0]
         assert_greater(hist[i], 40)
 
+def test_minsize():
+    # single-channel:
+    img = data.coins()[20:168,0:128]
+    for min_size in np.arange(10, 100, 10):
+        segments = felzenszwalb(img, min_size=min_size, sigma=3)
+        counts = np.bincount(segments.ravel())
+        # actually want to test greater or equal.
+        assert_greater(counts.min() + 1, min_size)
+    # multi-channel:
+    coffee = data.coffee()[::4, ::4]
+    for min_size in np.arange(10, 100, 10):
+        segments = felzenszwalb(coffee, min_size=min_size, sigma=3)
+        counts = np.bincount(segments.ravel())
+        # actually want to test greater or equal.
+        # the construction doesn't guarantee min_size is respected
+        # after intersecting the sementations for the colors
+        assert_greater(np.mean(counts) + 1, min_size)
+
 
 def test_color():
     # very weak tests. This algorithm is pretty unstable.