diff --git a/skimage/feature/_hog.py b/skimage/feature/_hog.py
index 7fafc10b..51d0d060 100644
--- a/skimage/feature/_hog.py
+++ b/skimage/feature/_hog.py
@@ -1,5 +1,4 @@
 import numpy as np
-from scipy import sqrt, pi, arctan2, cos, sin
 from scipy.ndimage import uniform_filter
 from .._shared.utils import assert_nD
 from . import _hoghistogram
@@ -63,7 +62,7 @@ def hog(image, orientations=9, pixels_per_cell=(8, 8),
     assert_nD(image, 2)
 
     if normalise:
-        image = sqrt(image)
+        image = np.sqrt(image)
 
     """
     The second stage computes first order image gradients. These capture
@@ -104,8 +103,8 @@ def hog(image, orientations=9, pixels_per_cell=(8, 8),
     cell are used to vote into the orientation histogram.
     """
 
-    magnitude = sqrt(gx ** 2 + gy ** 2)
-    orientation = arctan2(gy, gx) * (180 / pi) % 180
+    magnitude = np.hypot(gx, gy)
+    orientation = np.arctan2(gy, gx) * (180 / np.pi) % 180
 
     sy, sx = image.shape
     cx, cy = pixels_per_cell
@@ -132,8 +131,8 @@ def hog(image, orientations=9, pixels_per_cell=(8, 8),
             for y in range(n_cellsy):
                 for o in range(orientations):
                     centre = tuple([y * cy + cy // 2, x * cx + cx // 2])
-                    dx = radius * cos(float(o) / orientations * np.pi)
-                    dy = radius * sin(float(o) / orientations * np.pi)
+                    dx = radius * np.cos(float(o) / orientations * np.pi)
+                    dy = radius * np.sin(float(o) / orientations * np.pi)
                     rr, cc = draw.line(int(centre[0] - dx),
                                        int(centre[1] + dy),
                                        int(centre[0] + dx),
@@ -164,7 +163,7 @@ def hog(image, orientations=9, pixels_per_cell=(8, 8),
         for y in range(n_blocksy):
             block = orientation_histogram[y:y + by, x:x + bx, :]
             eps = 1e-5
-            normalised_blocks[y, x, :] = block / sqrt(block.sum() ** 2 + eps)
+            normalised_blocks[y, x, :] = block / np.sqrt(block.sum() ** 2 + eps)
 
     """
     The final step collects the HOG descriptors from all blocks of a dense
diff --git a/skimage/feature/_hoghistogram.pyx b/skimage/feature/_hoghistogram.pyx
index 42df311f..77c63f4b 100644
--- a/skimage/feature/_hoghistogram.pyx
+++ b/skimage/feature/_hoghistogram.pyx
@@ -15,25 +15,25 @@ cdef float CellHog(np.ndarray[np.float64_t, ndim=2] magnitude,
     Parameters
     ----------
     magnitude : ndarray
-        Coordinate to be clipped.
+        The gradient magnitudes of the pixels.
     orientation : ndarray
-        The lower bound.
+        Lookup table for orientations.
     ori1 : float
-        The higher bound.
+        Orientation range start.
     ori2 : float
-        The higher bound.
+        Orientation range end.
     cx : int
-        The higher bound.
+        Pixels per cell (x).
     cy : int
-        The higher bound.
+        Pixels per cell (y).
     xi : int
-        The higher bound.
+        Block index (x).
     yi : int
-        The higher bound.
+        Block index (y).
     sx : int
-        The higher bound.
+        Image size (x).
     sy : int
-        The higher bound.
+        Image size (y).
 
     Returns
     -------
@@ -58,35 +58,35 @@ cdef float CellHog(np.ndarray[np.float64_t, ndim=2] magnitude,
 
 def HogHistograms(np.ndarray[np.float64_t, ndim=2] gx, 
     np.ndarray[np.float64_t, ndim=2] gy, 
-    int cx, int cy, #Pixels per cell
-    int sx, int sy, #Image size
+    int cx, int cy, 
+    int sx, int sy, 
     int n_cellsx, int n_cellsy, 
     int visualise, int orientations, 
     np.ndarray[np.float64_t, ndim=3] orientation_histogram):
-    """HogHistograms
+    """Extract Histogram of Oriented Gradients (HOG) for a given image.
 
     Parameters
     ----------
     gx : ndarray
-        Coordinate to be clipped.
+        First order image gradients (x).
     gy : ndarray
-        The lower bound.
+        First order image gradients (y).
     cx : int
-        The higher bound.
+        Pixels per cell (x).
     cy : int
-        The higher bound.
+        Pixels per cell (y).
     sx : int
-        The higher bound.
+        Image size (x).
     sy : int
-        The higher bound.
+        Image size (y).
     n_cellsx : int
-        The higher bound.
+        Number of cells (x).
     n_cellsy : int
-        The higher bound.
+        Number of cells (y).
     visualise : int
-        The higher bound.
+        Also return an image of the HOG.
     orientations : int
-        The higher bound.
+        Number of orientation bins.
     orientation_histogram : ndarray
         The histogram to fill.
     """