Help with find_boundaries or watershed
Moin moin, I'm using the find_boundaries function for separating labeled regions (separated by the watershed). Afterward I set the as True returned pixels to 0. This works quit well but since my objects (labeled regions) can be small (couple pixels) and seldom stick together I would like to limit the find_boundaries to connected regions. Is there a easy way to only find the boundaries not connected to 0. Or tell the watershed function to leave a space between separate objects? Example: import numpy as np from scipy import ndimage as nd from skimage.segmentation.boundaries import find_boundaries a = np.zeros((15,15)) a[4:9,4:9]=1 a[6:10,7:12]=2 a[1:3,1:3]=3 print "a" print a.astype(int) a[find_boundaries(a)]=0 print a.astype(int) Output: [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 3 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 3 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 1 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 1 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 1 1 1 2 2 2 2 2 0 0 0] [0 0 0 0 1 1 1 2 2 2 2 2 0 0 0] [0 0 0 0 1 1 1 2 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 2 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]] After find_boundaries: [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 0 1 1 0 0 0 0 0 0 0 0] [0 0 0 0 0 1 1 0 2 2 2 2 0 0 0] [0 0 0 0 0 1 1 0 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]] As you can see, the region 3 is almost gone. Thanks Siggi
Hi Siggi, How about replacing
a[find_boundaries(a)]=0
with
s = nd.generate_binary_structure(2, 1) a[find_boundaries(a) * (nd.grey_erosion(a, footprint=s) != 0)] = 0
? That will leave you with big objects that are separated with respect to connectivity=1, and objects that aren't touching will not be "reduced". The output for your example is:
print a.astype(int) [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 3 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 3 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 1 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 1 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 1 1 1 0 0 2 2 2 0 0 0] [0 0 0 0 1 1 1 0 2 2 2 2 0 0 0] [0 0 0 0 1 1 1 0 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 2 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]
Juan. On Sat, Nov 24, 2012 at 3:13 AM, Sigmund <siggin@gmail.com> wrote:
Moin moin,
I'm using the find_boundaries function for separating labeled regions (separated by the watershed). Afterward I set the as True returned pixels to 0. This works quit well but since my objects (labeled regions) can be small (couple pixels) and seldom stick together I would like to limit the find_boundaries to connected regions. Is there a easy way to only find the boundaries not connected to 0. Or tell the watershed function to leave a space between separate objects?
Example: import numpy as np from scipy import ndimage as nd from skimage.segmentation.boundaries import find_boundaries a = np.zeros((15,15)) a[4:9,4:9]=1 a[6:10,7:12]=2 a[1:3,1:3]=3 print "a" print a.astype(int) a[find_boundaries(a)]=0 print a.astype(int)
Output:
[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 3 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 3 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 1 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 1 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 1 1 1 2 2 2 2 2 0 0 0] [0 0 0 0 1 1 1 2 2 2 2 2 0 0 0] [0 0 0 0 1 1 1 2 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 2 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]] After find_boundaries: [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 3 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 1 1 1 1 0 0 0 0 0 0] [0 0 0 0 0 1 1 0 0 0 0 0 0 0 0] [0 0 0 0 0 1 1 0 2 2 2 2 0 0 0] [0 0 0 0 0 1 1 0 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 2 2 2 2 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]
As you can see, the region 3 is almost gone.
Thanks Siggi
--
participants (2)
-
Juan Nunez-Iglesias -
Sigmund