Harmonize all ndimage usage across the library

Only two forms remain in use:

- `from scipy import ndimage as ndi`
- `from scipy.ndimage import function`

skimage/segmentation/_felzenszwalb_cy.pyx

@@ -3,7 +3,7 @@
 #cython: nonecheck=False
 #cython: wraparound=False
 import numpy as np
-import scipy
+from scipy import ndimage as ndi
 
 cimport cython
 cimport numpy as cnp
@@ -47,7 +47,7 @@ def _felzenszwalb_grey(image, double scale=1, sigma=0.8,
 
     # rescale scale to behave like in reference implementation
     scale = float(scale) / 255.
-    image = scipy.ndimage.gaussian_filter(image, sigma=sigma)
+    image = ndi.gaussian_filter(image, sigma=sigma)
 
     # compute edge weights in 8 connectivity:
     right_cost = np.abs((image[1:, :] - image[:-1, :]))

skimage/segmentation/_quickshift.pyx

@@ -3,7 +3,7 @@
 #cython: nonecheck=False
 #cython: wraparound=False
 import numpy as np
-from scipy import ndimage
+from scipy import ndimage as ndi
 from itertools import product
 
 cimport numpy as cnp
@@ -69,7 +69,7 @@ def quickshift(image, ratio=1., float kernel_size=5, max_dist=10,
             ValueError("Only RGB images can be converted to Lab space.")
         image = rgb2lab(image)
 
-    image = ndimage.gaussian_filter(img_as_float(image), [sigma, sigma, 0])
+    image = ndi.gaussian_filter(img_as_float(image), [sigma, sigma, 0])
     cdef cnp.ndarray[dtype=cnp.float_t, ndim=3, mode="c"] image_c \
             = np.ascontiguousarray(image) * ratio
 

skimage/segmentation/boundaries.py

@@ -1,7 +1,7 @@
 from __future__ import division
 
 import numpy as np
-from scipy import ndimage as nd
+from scipy import ndimage as ndi
 from ..morphology import dilation, erosion, square
 from ..util import img_as_float, view_as_windows, pad
 from ..color import gray2rgb
@@ -146,7 +146,7 @@ def find_boundaries(label_img, connectivity=1, mode='thick', background=0):
            [0, 0, 0, 0, 0, 0, 0]], dtype=uint8)
     """
     ndim = label_img.ndim
-    selem = nd.generate_binary_structure(ndim, connectivity)
+    selem = ndi.generate_binary_structure(ndim, connectivity)
     if mode != 'subpixel':
         boundaries = dilation(label_img, selem) != erosion(label_img, selem)
         if mode == 'inner':
@@ -155,7 +155,7 @@ def find_boundaries(label_img, connectivity=1, mode='thick', background=0):
         elif mode == 'outer':
             max_label = np.iinfo(label_img.dtype).max
             background_image = (label_img == background)
-            selem = nd.generate_binary_structure(ndim, ndim)
+            selem = ndi.generate_binary_structure(ndim, ndim)
             inverted_background = np.array(label_img, copy=True)
             inverted_background[background_image] = max_label
             adjacent_objects = ((dilation(label_img, selem) !=
@@ -205,10 +205,10 @@ def mark_boundaries(image, label_img, color=(1, 1, 0),
     if mode == 'subpixel':
         # Here, we want to interpose an extra line of pixels between
         # each original line - except for the last axis which holds
-        # the RGB information. ``nd.zoom`` then performs the (cubic)
+        # the RGB information. ``ndi.zoom`` then performs the (cubic)
         # interpolation, filling in the values of the interposed pixels
-        marked = nd.zoom(marked, [2 - 1/s for s in marked.shape[:-1]] + [1],
-                         mode='reflect')
+        marked = ndi.zoom(marked, [2 - 1/s for s in marked.shape[:-1]] + [1],
+                          mode='reflect')
     boundaries = find_boundaries(label_img, mode=mode,
                                  background=background_label)
     if outline_color is not None:

skimage/segmentation/random_walker_segmentation.py

@@ -10,7 +10,7 @@ significantly the performance.
 
 import warnings
 import numpy as np
-from scipy import sparse, ndimage
+from scipy import sparse, ndimage as ndi
 
 # executive summary for next code block: try to import umfpack from
 # scipy, but make sure not to raise a fuss if it fails since it's only
@@ -417,7 +417,7 @@ def random_walker(data, labels, beta=130, mode='bf', tol=1.e-3, copy=True,
     # If the array has pruned zones, be sure that no isolated pixels
     # exist between pruned zones (they could not be determined)
     if np.any(labels < 0):
-        filled = ndimage.binary_propagation(labels > 0, mask=labels >= 0)
+        filled = ndi.binary_propagation(labels > 0, mask=labels >= 0)
         labels[np.logical_and(np.logical_not(filled), labels == 0)] = -1
         del filled
     labels = np.atleast_3d(labels)

skimage/segmentation/slic_superpixels.py

@@ -2,7 +2,7 @@
 
 import collections as coll
 import numpy as np
-from scipy import ndimage
+from scipy import ndimage as ndi
 import warnings
 
 from ..util import img_as_float, regular_grid
@@ -139,7 +139,7 @@ def slic(image, n_segments=100, compactness=10., max_iter=10, sigma=0,
     if (sigma > 0).any():
         # add zero smoothing for multichannel dimension
         sigma = list(sigma) + [0]
-        image = ndimage.gaussian_filter(image, sigma)
+        image = ndi.gaussian_filter(image, sigma)
 
     if multichannel and (convert2lab or convert2lab is None):
         if image.shape[-1] != 3 and convert2lab: