def clean_labels(labels, rad, OPEN=3):
"""default cleaning. Fill holes, remove small and large objects and opening.
"""
labels = gray_fill_holes(labels)
labels = clear_border(labels, buffer_size=2)
labels = remove_small_objects(labels, rad[0]**2 * np.pi, connectivity=4)
antimask = remove_small_objects(labels, rad[1]**2 * np.pi, connectivity=4)
labels[antimask > 0] = False
labels = label(binary_opening(labels, np.ones((int(OPEN), int(OPEN))), iterations=1))
return labels
def agglomeration_seed(labels, img, MINSIZE=50, STEPS=100, FILSIZE=5, RATIO=0):
"""
MINSIZE: minimum area for a seed object. It can be smaller than actual objects.
STEPS: Larger it is, more resolution and computation
FILSIZE: argument for adaptive thresholding. Larger if capturing too much backrgound.
RATIO: argument for adaptive thresholding. Larger if capturing too much backrgound.
"""
seed = binary_erosion(labels, np.ones((3, 3)))
li = []
img = img.astype(np.float32)
mask = adaptive_thresh(img, RATIO, FILSIZE)
mask = binary_opening(mask, np.ones((3, 3)))
mask = remove_small_objects(mask, MINSIZE)
foreground = img[mask]
perclist = [
np.percentile(foreground, r) for r in np.linspace(0, 100, STEPS)
]
for _r in perclist:
thresed = remove_small_objects(img > _r, MINSIZE, connectivity=2) > 0
li.append(thresed.astype(np.uint16))
if seed is not None:
li.append((seed > 0).astype(np.uint16))
for l in li:
l[seed > 0] = 1
q = np.sum(np.dstack(li), axis=2)
p = label(q)
for ind in reversed(np.unique(q).tolist()):
c = seeding_separate(label(q >= ind), p)
w = watershed(q >= ind,
markers=c,
mask=(q >= ind),
watershed_line=True)
w[mask == 0] = 0
w = remove_small_objects(w, MINSIZE)
p = label(w, connectivity=2)
return p
def cytoplasm_levelset(labels, img, niter=20, dt=-0.5, thres=0.5):
""" Segment cytoplasm from supplied nuclear labels and probability map
Expand using level sets method from nuclei to membrane.
Uses an implementation of Level set method. See:
https://wiseodd.github.io/techblog/2016/11/05/levelset-method/
Args:
labels (numpy.ndarray): nuclear mask labels
img (numpy.ndarray): probability map
niter (int): step size to expand mask, number of iterations to run the levelset algorithm
dt (float): negative values for porgation, positive values for shrinking
thres (float): threshold of probability value to extend the nuclear mask to
Returns:
cytolabels (numpy.ndarray): cytoplasm mask labels
"""
from skimage.morphology import closing, disk, remove_small_holes
from utils.dlevel_set import dlevel_set
phi = labels.copy()
phi[labels == 0] = 1
phi[labels > 0] = -1
outlines = img.copy()
outlines = -outlines
outlines = outlines - outlines.min()
outlines = outlines/outlines.max()
mask = outlines < thres
phi = dlevel_set(phi, outlines, niter=niter, dt=dt, mask=mask)
cytolabels = label(remove_small_holes(label(phi < 0)))
cytolabels = closing(cytolabels, disk(3))
temp = cytolabels.copy()
temp[labels == 0] = 0
cytolabels = convert_labels(temp, labels, cytolabels)
return cytolabels
def watershed_divide(labels, regmax=10, min_size=100):
"""
divide objects in labels with watershed segmentation.
regmax:
min_size: objects smaller than this size will not be divided.
"""
from utils.subdetect_utils import watershed_labels
large_labels = remove_small_objects(labels, min_size, connectivity=4)
labels[large_labels > 0] = 0
ws_large = label(watershed_labels(large_labels, regmax))
ws_large += labels.max()
ws_large[ws_large == labels.max()] = 0
return labels + ws_large
def _wd(labels0, labels, img0, img1):
labels1 = -labels.copy()
rps0 = regionprops(labels0, img0)
from subdetect_operation import watershed_divide # DO NOT MOVE IT
from utils.track_utils import _find_match
untracked_labels = labels1.copy()
untracked_labels[untracked_labels < 0] = 0
wshed_labels = watershed_divide(untracked_labels,
regmax=REGMAX,
min_size=MIN_SIZE)
wshed_labels = label(wshed_labels)
store = regionprops(wshed_labels, img1)
good_cells = _find_match(rps0, store, DISPLACEMENT, MASSTHRES)
for gc in good_cells: # needed to reuse _update_labels_neck_cut
gccrds = gc.coords[0]
gc.raw_label = labels1[gccrds[0], gccrds[1]]
labels0, labels = _update_labels_neck_cut(labels0, labels1, good_cells)
labels0, labels = nn_closer(img0, img1, labels0, -labels, DISPLACEMENT,
MASSTHRES)
return labels0, labels, good_cells
def laplacian_levelset(labels, img, NITER=100, CURVE=3, PROP=-1):
return label(levelset_lap(img, labels, NITER, CURVE, PROP))