Hello Alex,
the ultimate erosion coincides with the local maxima of the distance map
(or the local minima of the inverse distance map).
In skimage, you can get these by:
from skimage.morphology import extrema local_maxima = extrema.local_maxima
<http://scikit-image.org/docs/dev/api/skimage.morphology.html#skimage.morphology.local_maxima>(distance_map)
peak_local_max is not equivalent to the UEP, but can be preferable in
certain situations.
I did not check your example in practice as I do not have the image, but
I was a bit confused by the inversions, as you start by applying the
threshold the other way around.
Best,
Thomas.
P.S.: The ImageJ description is not correct, which I have always found
disturbing. The Watershed algorithm is of course not limited to
applications to the inverse distance map.
On 10/5/17 3:22 PM, Alexandre Fioravante de Siqueira wrote:
Dear all,
I am trying to write a that mimics the behavior of ImageJ's flooding
watershed. This is how they describe it
(https://imagej.nih.gov/ij/docs/menus/process.html#watershed):
/Watershed
/
/Watershed segmentation is a way of automatically separating or
cutting apart particles that touch. It first calculates the Euclidian
distance map (EDM) and finds the ultimate eroded points (UEPs). It
then dilates each of the UEPs (the peaks or local maxima of the EDM)
as far as possible - either until the edge of the particle is reached,
or the edge of the region of another (growing) UEP. Watershed
segmentation works best for smooth convex objects that don't overlap
too much.
/
I tried to implement that (more or less), using the local minima (as
pointed in https://www.mathworks.com/help/images/ref/bwulterode.html).
Of course, there is something wrong :)
from scipy.ndimage import binary_fill_holes
from scipy.ndimage.morphology import distance_transform_edt
from skimage.feature import peak_local_max
from skimage.filters import threshold_otsu
from skimage.io <http://skimage.io> import imread
from skimage.measure import label
from skimage.morphology import disk, watershed
from skimage.util import invert
import matplotlib.pyplot as plt
image = imread('K90_incid4,5min_1.bmp')
img_bin = binary_fill_holes(image < threshold_otsu(image))
inv_bin = invert(img_bin)
inv_dist = distance_transform_edt(inv_bin)
inv_peak = peak_local_max(-inv_dist, indices=False)
markers = label(peak)
labels = watershed(-distance, markers, mask=img_bin, watershed_line=True)
plt.imshow(labels, cmap='nipy_spectral'), plt.show()
We don't have a function for UEPs in skimage yet, right?
How could I implement this flooding watershed?
Thank you very much. Kind regards,
Alex
--
--------------------------------------------------
Dr. Alexandre 'Jaguar' Fioravante de Siqueira
Grupo de Cronologia - Sala 13
Departamento de Raios Cósmicos e Cronologia - DRCC
Instituto de Física "Gleb Wataghin" - IFGW
Unicamp - Universidade Estadual de Campinas
Rua Sérgio Buarque de Holanda, 777
Cidade Universitária Zeferino Vaz - CEP 13083-859
Campinas - SP - Brazil
Phone: +55(19)3521-5362
Lattes curriculum: 3936721630855880
<http://lattes.cnpq.br/3936721630855880/>
ORCID: 0000-0003-1320-4347 <http://orcid.org/0000-0003-1320-4347>
Personal site: programmingscience.org <http://www.programmingscience.org/>
Github: alexandrejaguar <http://www.github.com/alexandrejaguar/>
Skype: alexandrejaguar
Twitter: alexdesiqueira <http://www.twitter.com/alexdesiqueira/>
--------------------------------------------------
_______________________________________________
scikit-image mailing list
scikit-image@python.org
https://mail.python.org/mailman/listinfo/scikit-image
--
Thomas Walter
27 rue des Acacias
75017 Paris
_______________________________________________
scikit-image mailing list
scikit-image@python.org
https://mail.python.org/mailman/listinfo/scikit-image
