DOC Answer Haralick FAQ about 16 bit images

Explain that haralick can require large amounts of memory for 16 bit
images.

docs/source/faq.rst

@@ -25,6 +25,24 @@ In any case, we all use the same data format: numpy arrays. It is very easy
 you want from each. All the packages use function based interfaces which make
 it easy to mix-and-match.
 
+I ran out of memory computing Haralick features on 16 bit images. Is it not
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+supported?
+~~~~~~~~~~~
+
+Yes, it is. However, in order to compute Haralick features, first a
+cooccurrence matrix is computed. This matrix has the size ``(ngrey, ngrey)``
+where ``ngrey`` is the largest grey value in the input. Thus, if your image has
+a very high dynamic range, you may not have the resources to compute the
+matrix.
+
+It is often a good idea to contrast stretch your images anyway. For example,
+using the following code, stretches your images to the 0-255 range::
+
+    im_stretched = mh.stretch(im)
+    features = mh.features.haralic(im_stretched)
+
+
 What are the parameters to Local Binary Patterns?
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~