Download as
Browse Code »

Commit 09915b7f1fe2b89823044c02c32911165671163d

Authored by Thiago Franco de Moraes
1 parent a8e514b4
Exists in vtk6

Watershed using vtk6

Showing 2 changed files with 54 additions and 47 deletions   Show diff stats
invesalius/data/styles.py
  Diff comments   View file @09915b7
... ... @@ -59,10 +59,14 @@ WATERSHED_OPERATIONS = {_("Erase"): BRUSH_ERASE,
59 59 _("Background"): BRUSH_BACKGROUND,}
60 60  
61 61 def get_LUT_value(data, window, level):
62   - return np.piecewise(data,
63   - [data <= (level - 0.5 - (window-1)/2),
64   - data > (level - 0.5 + (window-1)/2)],
65   - [0, window, lambda data: ((data - (level - 0.5))/(window-1) + 0.5)*(window)])
  62 + shape = data.shape
  63 + data_ = data.ravel()
  64 + data = np.piecewise(data_,
  65 + [data_ <= (level - 0.5 - (window-1)/2),
  66 + data_ > (level - 0.5 + (window-1)/2)],
  67 + [0, window, lambda data_: ((data_ - (level - 0.5))/(window-1) + 0.5)*(window)])
  68 + data.shape = shape
  69 + return data
66 70  
67 71 class BaseImageInteractorStyle(vtk.vtkInteractorStyleImage):
68 72 def __init__(self, viewer):
... ...
invesalius/data/watershed_process.py
  Diff comments   View file @09915b7
1   -import numpy as np
2   -from scipy import ndimage
3   -from scipy.ndimage import watershed_ift, generate_binary_structure
4   -from skimage.morphology import watershed
5   -from skimage import filter
6   -
7   -
8   -def get_LUT_value(data, window, level):
9   - return np.piecewise(data,
10   - [data <= (level - 0.5 - (window-1)/2),
11   - data > (level - 0.5 + (window-1)/2)],
12   - [0, 255, lambda data: ((data - (level - 0.5))/(window-1) + 0.5)*(255-0)])
13   -
14   -
15   -
16   -def do_watershed(image, markers, tfile, shape, bstruct, algorithm, mg_size, use_ww_wl, wl, ww, q):
17   - mask = np.memmap(tfile, shape=shape, dtype='uint8', mode='r+')
18   -
19   - if use_ww_wl:
20   - if algorithm == 'Watershed':
21   - tmp_image = ndimage.morphological_gradient(
22   - get_LUT_value(image, ww, wl).astype('uint16'),
23   - mg_size)
24   - tmp_mask = watershed(tmp_image, markers.astype('int16'), bstruct)
25   - else:
26   - tmp_image = get_LUT_value(image, ww, wl).astype('uint16')
27   - #tmp_image = ndimage.gaussian_filter(tmp_image, self.config.mg_size)
28   - #tmp_image = ndimage.morphological_gradient(
29   - #get_LUT_value(image, ww, wl).astype('uint16'),
30   - #self.config.mg_size)
31   - tmp_mask = watershed_ift(tmp_image, markers.astype('int16'), bstruct)
32   - else:
33   - if algorithm == 'Watershed':
34   - tmp_image = ndimage.morphological_gradient((image - image.min()).astype('uint16'), mg_size)
35   - tmp_mask = watershed(tmp_image, markers.astype('int16'), bstruct)
36   - else:
37   - tmp_image = (image - image.min()).astype('uint16')
38   - #tmp_image = ndimage.gaussian_filter(tmp_image, self.config.mg_size)
39   - #tmp_image = ndimage.morphological_gradient((image - image.min()).astype('uint16'), self.config.mg_size)
40   - tmp_mask = watershed_ift(tmp_image, markers.astype('int8'), bstruct)
41   - mask[:] = tmp_mask
42   - mask.flush()
43   - q.put(1)
  1 +import numpy as np
  2 +from scipy import ndimage
  3 +from scipy.ndimage import watershed_ift, generate_binary_structure
  4 +from skimage.morphology import watershed
  5 +from skimage import filter
  6 +
  7 +
  8 +def get_LUT_value(data, window, level):
  9 + shape = data.shape
  10 + data_ = data.ravel()
  11 + data = np.piecewise(data_,
  12 + [data_ <= (level - 0.5 - (window-1)/2),
  13 + data_ > (level - 0.5 + (window-1)/2)],
  14 + [0, window, lambda data_: ((data_ - (level - 0.5))/(window-1) + 0.5)*(window)])
  15 + data.shape = shape
  16 + return data
  17 +
  18 +
  19 +def do_watershed(image, markers, tfile, shape, bstruct, algorithm, mg_size, use_ww_wl, wl, ww, q):
  20 + mask = np.memmap(tfile, shape=shape, dtype='uint8', mode='r+')
  21 +
  22 + if use_ww_wl:
  23 + if algorithm == 'Watershed':
  24 + tmp_image = ndimage.morphological_gradient(
  25 + get_LUT_value(image, ww, wl).astype('uint16'),
  26 + mg_size)
  27 + tmp_mask = watershed(tmp_image, markers.astype('int16'), bstruct)
  28 + else:
  29 + tmp_image = get_LUT_value(image, ww, wl).astype('uint16')
  30 + #tmp_image = ndimage.gaussian_filter(tmp_image, self.config.mg_size)
  31 + #tmp_image = ndimage.morphological_gradient(
  32 + #get_LUT_value(image, ww, wl).astype('uint16'),
  33 + #self.config.mg_size)
  34 + tmp_mask = watershed_ift(tmp_image, markers.astype('int16'), bstruct)
  35 + else:
  36 + if algorithm == 'Watershed':
  37 + tmp_image = ndimage.morphological_gradient((image - image.min()).astype('uint16'), mg_size)
  38 + tmp_mask = watershed(tmp_image, markers.astype('int16'), bstruct)
  39 + else:
  40 + tmp_image = (image - image.min()).astype('uint16')
  41 + #tmp_image = ndimage.gaussian_filter(tmp_image, self.config.mg_size)
  42 + #tmp_image = ndimage.morphological_gradient((image - image.min()).astype('uint16'), self.config.mg_size)
  43 + tmp_mask = watershed_ift(tmp_image, markers.astype('int8'), bstruct)
  44 + mask[:] = tmp_mask
  45 + mask.flush()
  46 + q.put(1)
... ...