def run(self, workspace):
for object_name in (self.objects_x.value, self.objects_y.value):
if object_name not in workspace.object_set.object_names:
raise ValueError(
"The %s objects are missing from the pipeline." %
object_name)
objects_x = workspace.object_set.get_objects(self.objects_x.value)
objects_y = workspace.object_set.get_objects(self.objects_y.value)
assert (objects_x.shape == objects_y.shape
), "Objects sets must have the same dimensions"
labels_x = objects_x.segmented.copy()
labels_y = objects_y.segmented.copy()
output = self.combine_arrays(labels_x, labels_y)
output_labels = skimage.morphology.label(output)
output_objects = Objects()
output_objects.segmented = output_labels
workspace.object_set.add_objects(output_objects,
self.output_object.value)
if self.show_window:
workspace.display_data.input_object_x_name = self.objects_x.value
workspace.display_data.input_object_x = objects_x.segmented
workspace.display_data.input_object_y_name = self.objects_y.value
workspace.display_data.input_object_y = objects_y.segmented
workspace.display_data.output_object_name = self.output_object.value
workspace.display_data.output_object = output_objects.segmented
def run(self, workspace):
"""Find the outlines on the current image set
workspace - The workspace contains
pipeline - instance of cpp for this run
image_set - the images in the image set being processed
object_set - the objects (labeled masks) in this image set
measurements - the measurements for this run
frame - the parent frame to whatever frame is created. None means don't draw.
"""
gridding = workspace.get_grid(self.grid_name.value)
if self.shape_choice == SHAPE_RECTANGLE:
labels = self.run_rectangle(workspace, gridding)
elif self.shape_choice == SHAPE_CIRCLE_FORCED:
labels = self.run_forced_circle(workspace, gridding)
elif self.shape_choice == SHAPE_CIRCLE_NATURAL:
labels = self.run_natural_circle(workspace, gridding)
elif self.shape_choice == SHAPE_NATURAL:
labels = self.run_natural(workspace, gridding)
objects = Objects()
objects.segmented = labels
object_count = gridding.rows * gridding.columns
workspace.object_set.add_objects(objects,
self.output_objects_name.value)
add_object_location_measurements(workspace.measurements,
self.output_objects_name.value,
labels, object_count)
add_object_count_measurements(workspace.measurements,
self.output_objects_name.value,
object_count)
if self.show_window:
workspace.display_data.gridding = gridding
workspace.display_data.labels = labels
def run(self, workspace):
x_name = self.x_name.value
y_name = self.y_name.value
objects = workspace.object_set
x = objects.get_objects(x_name)
dimensions = x.dimensions
x_data = x.segmented
if self.method.value == "Dimensions":
y_data = resize(x_data, (self.height.value, self.width.value))
elif self.method.value == "Match Image":
target_image = workspace.image_set.get_image(
self.specific_image.value)
if target_image.volumetric:
tgt_height, tgt_width = target_image.pixel_data.shape[1:3]
else:
tgt_height, tgt_width = target_image.pixel_data.shape[:2]
size = (tgt_height, tgt_width)
y_data = resize(x_data, size)
else:
y_data = rescale(x_data, self.factor.value)
y = Objects()
y.segmented = y_data
y.parent_image = x.parent_image
objects.add_objects(y, y_name)
self.add_measurements(workspace)
if self.show_window:
workspace.display_data.x_data = x_data
workspace.display_data.y_data = y_data
workspace.display_data.dimensions = dimensions
def run(self, workspace):
x_name = self.x_name.value
y_name = self.y_name.value
images = workspace.image_set
x = images.get_image(x_name)
dimensions = x.dimensions
x_data = x.pixel_data
args = (setting.value for setting in self.settings()[2:])
y_data = self.function(x_data, *args)
y = Objects()
y.segmented = y_data
y.parent_image = x.parent_image
objects = workspace.object_set
objects.add_objects(y, y_name)
self.add_measurements(workspace)
if self.show_window:
workspace.display_data.x_data = x_data
workspace.display_data.y_data = y_data
workspace.display_data.dimensions = dimensions
def run(self, workspace):
image_name = self.image_name.value
objects_name = self.objects_name.value
image = workspace.image_set.get_image(image_name)
pixel_data = image.pixel_data
labels = workspace.interaction_request(
self, pixel_data, workspace.measurements.image_set_number)
if labels is None:
# User cancelled. Soldier on as best we can.
workspace.cancel_request()
labels = numpy.zeros(pixel_data.shape[:2], int)
objects = Objects()
objects.segmented = labels
workspace.object_set.add_objects(objects, objects_name)
##################
#
# Add measurements
#
m = workspace.measurements
#
# The object count
#
object_count = numpy.max(labels)
add_object_count_measurements(m, objects_name, object_count)
#
# The object locations
#
add_object_location_measurements(m, objects_name, labels)
workspace.display_data.labels = labels
workspace.display_data.pixel_data = pixel_data
def test_segmented(self):
segmented = numpy.ones((224, 224), numpy.bool)
objects = Objects()
objects.segmented = segmented
numpy.testing.assert_array_equal(objects.segmented, segmented)
def test_masked(self):
x = numpy.zeros((224, 224, 3), numpy.float32)
mask = numpy.ones((224, 224), numpy.bool)
parent_image = Image(x, mask=mask)
objects = Objects()
objects.segmented = mask
objects.parent_image = parent_image
numpy.testing.assert_array_equal(objects.masked, mask)
def run(self, workspace):
for object_name in (self.objects_x.value, self.objects_y.value):
if object_name not in workspace.object_set.object_names:
raise ValueError(
"The %s objects are missing from the pipeline." %
object_name)
objects_x = workspace.object_set.get_objects(self.objects_x.value)
objects_y = workspace.object_set.get_objects(self.objects_y.value)
dimensions = objects_x.dimensions
assert (objects_x.shape == objects_y.shape
), "Objects sets must have the same dimensions"
labels_x = objects_x.segmented.copy().astype("uint16")
labels_y = objects_y.segmented.copy().astype("uint16")
output = combineobjects(self.merge_method.value, labels_x, labels_y,
dimensions)
output_labels = skimage.morphology.label(output)
output_objects = Objects()
output_objects.segmented = output_labels
workspace.object_set.add_objects(output_objects,
self.output_object.value)
m = workspace.measurements
object_count = numpy.max(output_labels)
add_object_count_measurements(m, self.output_object.value,
object_count)
add_object_location_measurements(m, self.output_object.value,
output_labels)
if self.show_window:
workspace.display_data.input_object_x_name = self.objects_x.value
workspace.display_data.input_object_x = objects_x.segmented
workspace.display_data.input_object_y_name = self.objects_y.value
workspace.display_data.input_object_y = objects_y.segmented
workspace.display_data.output_object_name = self.output_object.value
workspace.display_data.output_object = output_objects.segmented
workspace.display_data.dimensions = dimensions
def run(self, workspace):
x_name = self.x_name.value
y_name = self.y_name.value
objects = workspace.object_set
x = objects.get_objects(x_name)
dimensions = x.dimensions
x_data = x.segmented
if self.method.value == "Dimensions":
if x_data.ndim == 3:
size = (self.planes.value, self.height.value, self.width.value)
else:
size = (self.height.value, self.width.value)
y_data = resize(x_data, size)
elif self.method.value == "Match Image":
target_image = workspace.image_set.get_image(
self.specific_image.value)
if target_image.volumetric:
size = target_image.pixel_data.shape[:3]
else:
size = target_image.pixel_data.shape[:2]
y_data = resize(x_data, size)
else:
if x_data.ndim == 3:
size = (self.factor_z.value, self.factor_y.value,
self.factor_x.value)
else:
size = (self.factor_y.value, self.factor_x.value)
y_data = rescale(x_data, size)
y = Objects()
y.segmented = y_data
objects.add_objects(y, y_name)
self.add_measurements(workspace)
if self.show_window:
workspace.display_data.x_data = x_data
workspace.display_data.y_data = y_data
workspace.display_data.dimensions = dimensions
def run(self, workspace):
x_name = self.x_name.value
y_name = self.y_name.value
objects = workspace.object_set
x = objects.get_objects(x_name)
dimensions = x.dimensions
x_data = x.segmented
props = skimage.measure.regionprops(
x_data
) # , properties=('label', 'centroid'))
y_data = numpy.zeros_like(x_data)
for region in props:
label = region.label
binary = x_data == label
eroded = cellprofiler.utilities.morphology.binary_erosion(
binary, self.structuring_element.value
)
y_data[eroded] = label
if self.preserve_midpoints.value:
if label not in y_data:
midpoint = scipy.ndimage.morphology.distance_transform_edt(binary)
y_data[midpoint == numpy.max(midpoint)] = label
if self.relabel_objects.value:
y_data = skimage.morphology.label(y_data)
y = Objects()
y.segmented = y_data
y.parent_image = x.parent_image
objects.add_objects(y, y_name)
self.add_measurements(workspace)
if self.show_window:
workspace.display_data.x_data = x_data
workspace.display_data.y_data = y_data
workspace.display_data.dimensions = dimensions
def run(self, workspace):
"""Run the module on the current data set
workspace - has the current image set, object set, measurements
and the parent frame for the application if the module
is allowed to display. If the module should not display,
workspace.frame is None.
"""
#
# The object set holds "objects". Each of these is a container
# for holding up to three kinds of image labels.
#
object_set = workspace.object_set
#
# Get the primary objects (the centers to be removed).
# Get the string value out of primary_object_name.
#
primary_objects = object_set.get_objects(
self.primary_objects_name.value)
#
# Get the cleaned-up labels image
#
primary_labels = primary_objects.segmented
#
# Do the same with the secondary object
secondary_objects = object_set.get_objects(
self.secondary_objects_name.value)
secondary_labels = secondary_objects.segmented
#
# If one of the two label images is smaller than the other, we
# try to find the cropping mask and we apply that mask to the larger
#
try:
if any([
p_size < s_size for p_size, s_size in zip(
primary_labels.shape, secondary_labels.shape)
]):
#
# Look for a cropping mask associated with the primary_labels
# and apply that mask to resize the secondary labels
#
secondary_labels = primary_objects.crop_image_similarly(
secondary_labels)
tertiary_image = primary_objects.parent_image
elif any([
p_size > s_size for p_size, s_size in zip(
primary_labels.shape, secondary_labels.shape)
]):
primary_labels = secondary_objects.crop_image_similarly(
primary_labels)
tertiary_image = secondary_objects.parent_image
elif secondary_objects.parent_image is not None:
tertiary_image = secondary_objects.parent_image
else:
tertiary_image = primary_objects.parent_image
except ValueError:
# No suitable cropping - resize all to fit the secondary
# labels which are the most critical.
#
primary_labels, _ = size_similarly(secondary_labels,
primary_labels)
if secondary_objects.parent_image is not None:
tertiary_image = secondary_objects.parent_image
else:
tertiary_image = primary_objects.parent_image
if tertiary_image is not None:
tertiary_image, _ = size_similarly(secondary_labels,
tertiary_image)
# If size/shape differences were too extreme, raise an error.
if primary_labels.shape != secondary_labels.shape:
raise ValueError(
"This module requires that the object sets have matching widths and matching heights.\n"
"The %s and %s objects do not (%s vs %s).\n"
"If they are paired correctly you may want to use the ResizeObjects module "
"to make them the same size." % (
self.secondary_objects_name,
self.primary_objects_name,
secondary_labels.shape,
primary_labels.shape,
))
#
# Find the outlines of the primary image and use this to shrink the
# primary image by one. This guarantees that there is something left
# of the secondary image after subtraction
#
primary_outline = outline(primary_labels)
tertiary_labels = secondary_labels.copy()
if self.shrink_primary:
primary_mask = numpy.logical_or(primary_labels == 0,
primary_outline)
else:
primary_mask = primary_labels == 0
tertiary_labels[primary_mask == False] = 0
#
# Get the outlines of the tertiary image
#
tertiary_outlines = outline(tertiary_labels) != 0
#
# Make the tertiary objects container
#
tertiary_objects = Objects()
tertiary_objects.segmented = tertiary_labels
tertiary_objects.parent_image = tertiary_image
#
# Relate tertiary objects to their parents & record
#
child_count_of_secondary, secondary_parents = secondary_objects.relate_children(
tertiary_objects)
if self.shrink_primary:
child_count_of_primary, primary_parents = primary_objects.relate_children(
tertiary_objects)
else:
# Primary and tertiary don't overlap.
# Establish overlap between primary and secondary and commute
_, secondary_of_primary = secondary_objects.relate_children(
primary_objects)
mask = secondary_of_primary != 0
child_count_of_primary = numpy.zeros(mask.shape, int)
child_count_of_primary[mask] = child_count_of_secondary[
secondary_of_primary[mask] - 1]
primary_parents = numpy.zeros(secondary_parents.shape,
secondary_parents.dtype)
primary_of_secondary = numpy.zeros(secondary_objects.count + 1,
int)
primary_of_secondary[secondary_of_primary] = numpy.arange(
1,
len(secondary_of_primary) + 1)
primary_of_secondary[0] = 0
primary_parents = primary_of_secondary[secondary_parents]
#
# Write out the objects
#
workspace.object_set.add_objects(tertiary_objects,
self.subregion_objects_name.value)
#
# Write out the measurements
#
m = workspace.measurements
#
# The parent/child associations
#
for parent_objects_name, parents_of, child_count, relationship in (
(
self.primary_objects_name,
primary_parents,
child_count_of_primary,
R_REMOVED,
),
(
self.secondary_objects_name,
secondary_parents,
child_count_of_secondary,
R_PARENT,
),
):
m.add_measurement(
self.subregion_objects_name.value,
FF_PARENT % parent_objects_name.value,
parents_of,
)
m.add_measurement(
parent_objects_name.value,
FF_CHILDREN_COUNT % self.subregion_objects_name.value,
child_count,
)
mask = parents_of != 0
image_number = numpy.ones(numpy.sum(mask),
int) * m.image_set_number
child_object_number = numpy.argwhere(mask).flatten() + 1
parent_object_number = parents_of[mask]
m.add_relate_measurement(
self.module_num,
relationship,
parent_objects_name.value,
self.subregion_objects_name.value,
image_number,
parent_object_number,
image_number,
child_object_number,
)
object_count = tertiary_objects.count
#
# The object count
#
add_object_count_measurements(workspace.measurements,
self.subregion_objects_name.value,
object_count)
#
# The object locations
#
add_object_location_measurements(workspace.measurements,
self.subregion_objects_name.value,
tertiary_labels)
if self.show_window:
workspace.display_data.primary_labels = primary_labels
workspace.display_data.secondary_labels = secondary_labels
workspace.display_data.tertiary_labels = tertiary_labels
workspace.display_data.tertiary_outlines = tertiary_outlines
def run(self, workspace):
images = workspace.image_set
x = images.get_image(self.x_name.value)
dimensions = x.dimensions
x_data = x.pixel_data
# Validate some settings
if self.model.value in (GREY_1, GREY_2) and x.multichannel:
raise ValueError(
"Color images are not supported by this model. Please provide greyscale images."
)
elif self.model.value == COLOR_1 and not x.multichannel:
raise ValueError(
"Greyscale images are not supported by this model. Please provide a color overlay."
)
if self.model.value != MODEL_CUSTOM:
if x.volumetric:
raise NotImplementedError(
"StarDist's inbuilt models do not currently support 3D images"
)
model = StarDist2D.from_pretrained(self.model.value)
else:
model_directory, model_name = os.path.split(
self.model_directory.get_absolute_path())
if x.volumetric:
from stardist.models import StarDist3D
model = StarDist3D(config=None,
basedir=model_directory,
name=model_name)
else:
model = StarDist2D(config=None,
basedir=model_directory,
name=model_name)
tiles = None
if self.tile_image.value:
tiles = []
if x.volumetric:
tiles += [1]
tiles += [self.n_tiles_x.value, self.n_tiles_y.value]
# Handle colour channels
tiles += [1] * max(0, x.pixel_data.ndim - len(tiles))
print(x.pixel_data.shape, x.pixel_data.ndim, tiles)
if not self.save_probabilities.value:
# Probabilities aren't wanted, things are simple
data = model.predict_instances(normalize(x.pixel_data),
return_predict=False,
n_tiles=tiles)
y_data = data[0]
else:
data, probs = model.predict_instances(normalize(x.pixel_data),
return_predict=True,
sparse=False,
n_tiles=tiles)
y_data = data[0]
# Scores aren't at the same resolution as the input image.
# We need to slightly resize to match the original image.
size_corrected = resize(probs[0], y_data.shape)
prob_image = Image(
size_corrected,
parent_image=x.parent_image,
convert=False,
dimensions=len(size_corrected.shape),
)
workspace.image_set.add(self.probabilities_name.value, prob_image)
if self.show_window:
workspace.display_data.probabilities = size_corrected
y = Objects()
y.segmented = y_data
y.parent_image = x.parent_image
objects = workspace.object_set
objects.add_objects(y, self.y_name.value)
self.add_measurements(workspace)
if self.show_window:
workspace.display_data.x_data = x_data
workspace.display_data.y_data = y_data
workspace.display_data.dimensions = dimensions
def run_on_objects(self, object_name, workspace):
"""Determine desired measurements and pass in object arrays for analysis"""
objects = workspace.get_objects(object_name)
# Don't analyze if there are no objects at all.
if len(objects.indices) == 0:
# No objects to process
self.measurements_without_objects(workspace, object_name)
return
# Determine which properties we're measuring.
if len(objects.shape) == 2:
desired_properties = [
"label",
"image",
"area",
"perimeter",
"bbox",
"bbox_area",
"major_axis_length",
"minor_axis_length",
"orientation",
"centroid",
"equivalent_diameter",
"extent",
"eccentricity",
"solidity",
"euler_number",
]
if self.calculate_advanced.value:
desired_properties += [
"inertia_tensor",
"inertia_tensor_eigvals",
"moments",
"moments_central",
"moments_hu",
"moments_normalized",
]
else:
desired_properties = [
"label",
"image",
"area",
"centroid",
"bbox",
"bbox_area",
"major_axis_length",
"minor_axis_length",
"extent",
"equivalent_diameter",
"euler_number",
]
if self.calculate_advanced.value:
desired_properties += [
"solidity",
]
# Check for overlapping object sets
if not objects.overlapping():
features_to_record = self.analyze_objects(objects,
desired_properties)
else:
# Objects are overlapping, process as single arrays
coords_array = objects.ijv
features_to_record = {}
for label in objects.indices:
omap = numpy.zeros(objects.shape)
ocoords = coords_array[coords_array[:, 2] == label, 0:2]
numpy.put(omap, numpy.ravel_multi_index(ocoords.T, omap.shape),
1)
tempobject = Objects()
tempobject.segmented = omap
buffer = self.analyze_objects(tempobject, desired_properties)
for f, m in buffer.items():
if f in features_to_record:
features_to_record[f] = numpy.concatenate(
(features_to_record[f], m))
else:
features_to_record[f] = m
for f, m in features_to_record.items():
self.record_measurement(workspace, object_name, f, m)
def run(self, workspace):
if self.mode.value != MODE_CUSTOM:
model = models.Cellpose(model_type='cyto' if self.mode.value
== MODE_CELLS else 'nuclei',
gpu=self.use_gpu.value)
else:
model_file = self.model_file_name.value
model_directory = self.model_directory.get_absolute_path()
model_path = os.path.join(model_directory, model_file)
model = models.CellposeModel(pretrained_model=model_path,
gpu=self.use_gpu.value)
x_name = self.x_name.value
y_name = self.y_name.value
images = workspace.image_set
x = images.get_image(x_name)
dimensions = x.dimensions
x_data = x.pixel_data
if x.multichannel:
raise ValueError(
"Color images are not currently supported. Please provide greyscale images."
)
if self.mode.value != "Nuclei" and self.supply_nuclei.value:
nuc_image = images.get_image(self.nuclei_image.value)
# CellPose expects RGB, we'll have a blank red channel, cells in green and nuclei in blue.
if x.volumetric:
x_data = numpy.stack(
(numpy.zeros_like(x_data), x_data, nuc_image.pixel_data),
axis=1)
else:
x_data = numpy.stack(
(numpy.zeros_like(x_data), x_data, nuc_image.pixel_data),
axis=-1)
channels = [2, 3]
else:
channels = [0, 0]
diam = self.expected_diameter.value if self.expected_diameter.value > 0 else None
try:
y_data, flows, *_ = model.eval(
x_data,
channels=channels,
diameter=diam,
net_avg=self.use_averaging.value,
do_3D=x.volumetric,
flow_threshold=self.flow_threshold.value,
cellprob_threshold=self.dist_threshold.value)
finally:
if self.use_gpu.value and model.torch:
# Try to clear some GPU memory for other worker processes.
try:
from torch import cuda
cuda.empty_cache()
except Exception as e:
print(
f"Unable to clear GPU memory. You may need to restart CellProfiler to change models. {e}"
)
y = Objects()
y.segmented = y_data
y.parent_image = x.parent_image
objects = workspace.object_set
objects.add_objects(y, y_name)
if self.save_probabilities.value:
# Flows come out sized relative to CellPose's inbuilt model size.
# We need to slightly resize to match the original image.
size_corrected = resize(flows[2], y_data.shape)
prob_image = Image(
size_corrected,
parent_image=x.parent_image,
convert=False,
dimensions=len(size_corrected.shape),
)
workspace.image_set.add(self.probabilities_name.value, prob_image)
if self.show_window:
workspace.display_data.probabilities = size_corrected
self.add_measurements(workspace)
if self.show_window:
if x.volumetric:
# Can't show CellPose-accepted colour images in 3D
workspace.display_data.x_data = x.pixel_data
else:
workspace.display_data.x_data = x_data
workspace.display_data.y_data = y_data
workspace.display_data.dimensions = dimensions
def run(self, workspace):
"""Run the module on an image set"""
object_name = self.object_name.value
remaining_object_name = self.remaining_objects.value
original_objects = workspace.object_set.get_objects(object_name)
if self.mask_choice == MC_IMAGE:
mask = workspace.image_set.get_image(
self.masking_image.value, must_be_binary=True
)
mask = mask.pixel_data
else:
masking_objects = workspace.object_set.get_objects(
self.masking_objects.value
)
mask = masking_objects.segmented > 0
if self.wants_inverted_mask:
mask = ~mask
#
# Load the labels
#
labels = original_objects.segmented.copy()
nobjects = numpy.max(labels)
#
# Resize the mask to cover the objects
#
mask, m1 = size_similarly(labels, mask)
mask[~m1] = False
#
# Apply the mask according to the overlap choice.
#
if nobjects == 0:
pass
elif self.overlap_choice == P_MASK:
labels = labels * mask
else:
pixel_counts = fixup_scipy_ndimage_result(
scipy.ndimage.sum(
mask, labels, numpy.arange(1, nobjects + 1, dtype=numpy.int32)
)
)
if self.overlap_choice == P_KEEP:
keep = pixel_counts > 0
else:
total_pixels = fixup_scipy_ndimage_result(
scipy.ndimage.sum(
numpy.ones(labels.shape),
labels,
numpy.arange(1, nobjects + 1, dtype=numpy.int32),
)
)
if self.overlap_choice == P_REMOVE:
keep = pixel_counts == total_pixels
elif self.overlap_choice == P_REMOVE_PERCENTAGE:
fraction = self.overlap_fraction.value
keep = pixel_counts / total_pixels >= fraction
else:
raise NotImplementedError(
"Unknown overlap-handling choice: %s", self.overlap_choice.value
)
keep = numpy.hstack(([False], keep))
labels[~keep[labels]] = 0
#
# Renumber the labels matrix if requested
#
if self.retain_or_renumber == R_RENUMBER:
unique_labels = numpy.unique(labels[labels != 0])
indexer = numpy.zeros(nobjects + 1, int)
indexer[unique_labels] = numpy.arange(1, len(unique_labels) + 1)
labels = indexer[labels]
parent_objects = unique_labels
else:
parent_objects = numpy.arange(1, nobjects + 1)
#
# Add the objects
#
remaining_objects = Objects()
remaining_objects.segmented = labels
remaining_objects.unedited_segmented = original_objects.unedited_segmented
workspace.object_set.add_objects(remaining_objects, remaining_object_name)
#
# Add measurements
#
m = workspace.measurements
m.add_measurement(
remaining_object_name, FF_PARENT % object_name, parent_objects,
)
if numpy.max(original_objects.segmented) == 0:
child_count = numpy.array([], int)
else:
child_count = fixup_scipy_ndimage_result(
scipy.ndimage.sum(
labels,
original_objects.segmented,
numpy.arange(1, nobjects + 1, dtype=numpy.int32),
)
)
child_count = (child_count > 0).astype(int)
m.add_measurement(
object_name, FF_CHILDREN_COUNT % remaining_object_name, child_count,
)
if self.retain_or_renumber == R_RETAIN:
remaining_object_count = nobjects
else:
remaining_object_count = len(unique_labels)
add_object_count_measurements(m, remaining_object_name, remaining_object_count)
add_object_location_measurements(m, remaining_object_name, labels)
#
# Save the input, mask and output images for display
#
if self.show_window:
workspace.display_data.original_labels = original_objects.segmented
workspace.display_data.final_labels = labels
workspace.display_data.mask = mask
def run(self, workspace):
objects_name = self.objects_name.value
objects = workspace.object_set.get_objects(objects_name)
assert isinstance(objects, Objects)
labels = objects.segmented
if self.relabel_option == OPTION_SPLIT:
output_labels, count = scipy.ndimage.label(
labels > 0, numpy.ones((3, 3), bool))
else:
if self.merge_option == UNIFY_DISTANCE:
mask = labels > 0
if self.distance_threshold.value > 0:
#
# Take the distance transform of the reverse of the mask
# and figure out what points are less than 1/2 of the
# distance from an object.
#
d = scipy.ndimage.distance_transform_edt(~mask)
mask = d < self.distance_threshold.value / 2 + 1
output_labels, count = scipy.ndimage.label(
mask, numpy.ones((3, 3), bool))
output_labels[labels == 0] = 0
if self.wants_image:
output_labels = self.filter_using_image(workspace, mask)
elif self.merge_option == UNIFY_PARENT:
parents_name = self.parent_object.value
parents_of = workspace.measurements[objects_name, "_".join(
(C_PARENT, parents_name))]
output_labels = labels.copy().astype(numpy.uint32)
output_labels[labels > 0] = parents_of[labels[labels > 0] - 1]
if self.merging_method == UM_CONVEX_HULL:
ch_pts, n_pts = centrosome.cpmorphology.convex_hull(
output_labels)
ijv = centrosome.cpmorphology.fill_convex_hulls(
ch_pts, n_pts)
output_labels[ijv[:, 0], ijv[:, 1]] = ijv[:, 2]
output_objects = Objects()
output_objects.segmented = output_labels
if objects.has_small_removed_segmented:
output_objects.small_removed_segmented = copy_labels(
objects.small_removed_segmented, output_labels)
if objects.has_unedited_segmented:
output_objects.unedited_segmented = copy_labels(
objects.unedited_segmented, output_labels)
output_objects.parent_image = objects.parent_image
workspace.object_set.add_objects(output_objects,
self.output_objects_name.value)
measurements = workspace.measurements
add_object_count_measurements(
measurements,
self.output_objects_name.value,
numpy.max(output_objects.segmented),
)
add_object_location_measurements(measurements,
self.output_objects_name.value,
output_objects.segmented)
#
# Relate the output objects to the input ones and record
# the relationship.
#
children_per_parent, parents_of_children = objects.relate_children(
output_objects)
measurements.add_measurement(
self.objects_name.value,
FF_CHILDREN_COUNT % self.output_objects_name.value,
children_per_parent,
)
measurements.add_measurement(
self.output_objects_name.value,
FF_PARENT % self.objects_name.value,
parents_of_children,
)
if self.show_window:
workspace.display_data.orig_labels = objects.segmented
workspace.display_data.output_labels = output_objects.segmented
if self.merge_option == UNIFY_PARENT:
workspace.display_data.parent_labels = workspace.object_set.get_objects(
self.parent_object.value).segmented
def run(self, workspace):
image_name = self.image_name.value
image = workspace.image_set.get_image(image_name,
must_be_grayscale=True)
workspace.display_data.statistics = []
img = image.pixel_data
mask = image.mask
objects = workspace.object_set.get_objects(self.x_name.value)
if img.shape != objects.shape:
raise ValueError(
"This module requires that the input image and object sets are the same size.\n"
"The %s image and %s objects are not (%s vs %s).\n"
"If they are paired correctly you may want to use the Resize, ResizeObjects or "
"Crop module(s) to make them the same size." % (
image_name,
self.x_name.value,
img.shape,
objects.shape,
))
global_threshold = None
if self.method == M_DISTANCE_N:
has_threshold = False
else:
thresholded_image, global_threshold, sigma = self._threshold_image(
image_name, workspace)
workspace.display_data.global_threshold = global_threshold
workspace.display_data.threshold_sigma = sigma
has_threshold = True
#
# Get the following labels:
# * all edited labels
# * labels touching the edge, including small removed
#
labels_in = objects.unedited_segmented.copy()
labels_touching_edge = numpy.hstack(
(labels_in[0, :], labels_in[-1, :], labels_in[:,
0], labels_in[:,
-1]))
labels_touching_edge = numpy.unique(labels_touching_edge)
is_touching = numpy.zeros(numpy.max(labels_in) + 1, bool)
is_touching[labels_touching_edge] = True
is_touching = is_touching[labels_in]
labels_in[(~is_touching) & (objects.segmented == 0)] = 0
#
# Stretch the input labels to match the image size. If there's no
# label matrix, then there's no label in that area.
#
if tuple(labels_in.shape) != tuple(img.shape):
tmp = numpy.zeros(img.shape, labels_in.dtype)
i_max = min(img.shape[0], labels_in.shape[0])
j_max = min(img.shape[1], labels_in.shape[1])
tmp[:i_max, :j_max] = labels_in[:i_max, :j_max]
labels_in = tmp
if self.method in (M_DISTANCE_B, M_DISTANCE_N):
if self.method == M_DISTANCE_N:
distances, (i, j) = scipy.ndimage.distance_transform_edt(
labels_in == 0, return_indices=True)
labels_out = numpy.zeros(labels_in.shape, int)
dilate_mask = distances <= self.distance_to_dilate.value
labels_out[dilate_mask] = labels_in[i[dilate_mask],
j[dilate_mask]]
else:
labels_out, distances = centrosome.propagate.propagate(
img, labels_in, thresholded_image, 1.0)
labels_out[distances > self.distance_to_dilate.value] = 0
labels_out[labels_in > 0] = labels_in[labels_in > 0]
if self.fill_holes:
label_mask = labels_out == 0
small_removed_segmented_out = centrosome.cpmorphology.fill_labeled_holes(
labels_out, mask=label_mask)
else:
small_removed_segmented_out = labels_out
#
# Create the final output labels by removing labels in the
# output matrix that are missing from the segmented image
#
segmented_labels = objects.segmented
segmented_out = self.filter_labels(small_removed_segmented_out,
objects, workspace)
elif self.method == M_PROPAGATION:
labels_out, distance = centrosome.propagate.propagate(
img, labels_in, thresholded_image,
self.regularization_factor.value)
if self.fill_holes:
label_mask = labels_out == 0
small_removed_segmented_out = centrosome.cpmorphology.fill_labeled_holes(
labels_out, mask=label_mask)
else:
small_removed_segmented_out = labels_out.copy()
segmented_out = self.filter_labels(small_removed_segmented_out,
objects, workspace)
elif self.method == M_WATERSHED_G:
#
# First, apply the sobel filter to the image (both horizontal
# and vertical). The filter measures gradient.
#
sobel_image = numpy.abs(scipy.ndimage.sobel(img))
#
# Combine the image mask and threshold to mask the watershed
#
watershed_mask = numpy.logical_or(thresholded_image, labels_in > 0)
watershed_mask = numpy.logical_and(watershed_mask, mask)
#
# Perform the first watershed
#
labels_out = skimage.segmentation.watershed(
connectivity=numpy.ones((3, 3), bool),
image=sobel_image,
markers=labels_in,
mask=watershed_mask,
)
if self.fill_holes:
label_mask = labels_out == 0
small_removed_segmented_out = centrosome.cpmorphology.fill_labeled_holes(
labels_out, mask=label_mask)
else:
small_removed_segmented_out = labels_out.copy()
segmented_out = self.filter_labels(small_removed_segmented_out,
objects, workspace)
elif self.method == M_WATERSHED_I:
#
# invert the image so that the maxima are filled first
# and the cells compete over what's close to the threshold
#
inverted_img = 1 - img
#
# Same as above, but perform the watershed on the original image
#
watershed_mask = numpy.logical_or(thresholded_image, labels_in > 0)
watershed_mask = numpy.logical_and(watershed_mask, mask)
#
# Perform the watershed
#
labels_out = skimage.segmentation.watershed(
connectivity=numpy.ones((3, 3), bool),
image=inverted_img,
markers=labels_in,
mask=watershed_mask,
)
if self.fill_holes:
label_mask = labels_out == 0
small_removed_segmented_out = centrosome.cpmorphology.fill_labeled_holes(
labels_out, mask=label_mask)
else:
small_removed_segmented_out = labels_out
segmented_out = self.filter_labels(small_removed_segmented_out,
objects, workspace)
if self.wants_discard_edge and self.wants_discard_primary:
#
# Make a new primary object
#
lookup = scipy.ndimage.maximum(
segmented_out,
objects.segmented,
list(range(numpy.max(objects.segmented) + 1)),
)
lookup = centrosome.cpmorphology.fixup_scipy_ndimage_result(lookup)
lookup[0] = 0
lookup[lookup != 0] = numpy.arange(numpy.sum(lookup != 0)) + 1
segmented_labels = lookup[objects.segmented]
segmented_out = lookup[segmented_out]
new_objects = Objects()
new_objects.segmented = segmented_labels
if objects.has_unedited_segmented:
new_objects.unedited_segmented = objects.unedited_segmented
if objects.has_small_removed_segmented:
new_objects.small_removed_segmented = objects.small_removed_segmented
new_objects.parent_image = objects.parent_image
#
# Add the objects to the object set
#
objects_out = Objects()
objects_out.unedited_segmented = small_removed_segmented_out
objects_out.small_removed_segmented = small_removed_segmented_out
objects_out.segmented = segmented_out
objects_out.parent_image = image
objname = self.y_name.value
workspace.object_set.add_objects(objects_out, objname)
object_count = numpy.max(segmented_out)
#
# Add measurements
#
measurements = workspace.measurements
super(IdentifySecondaryObjects, self).add_measurements(workspace)
#
# Relate the secondary objects to the primary ones and record
# the relationship.
#
children_per_parent, parents_of_children = objects.relate_children(
objects_out)
measurements.add_measurement(
self.x_name.value,
FF_CHILDREN_COUNT % objname,
children_per_parent,
)
measurements.add_measurement(
objname,
FF_PARENT % self.x_name.value,
parents_of_children,
)
image_numbers = (numpy.ones(len(parents_of_children), int) *
measurements.image_set_number)
mask = parents_of_children > 0
measurements.add_relate_measurement(
self.module_num,
R_PARENT,
self.x_name.value,
self.y_name.value,
image_numbers[mask],
parents_of_children[mask],
image_numbers[mask],
numpy.arange(1,
len(parents_of_children) + 1)[mask],
)
#
# If primary objects were created, add them
#
if self.wants_discard_edge and self.wants_discard_primary:
workspace.object_set.add_objects(
new_objects, self.new_primary_objects_name.value)
super(IdentifySecondaryObjects, self).add_measurements(
workspace,
input_object_name=self.x_name.value,
output_object_name=self.new_primary_objects_name.value,
)
children_per_parent, parents_of_children = new_objects.relate_children(
objects_out)
measurements.add_measurement(
self.new_primary_objects_name.value,
FF_CHILDREN_COUNT % objname,
children_per_parent,
)
measurements.add_measurement(
objname,
FF_PARENT % self.new_primary_objects_name.value,
parents_of_children,
)
if self.show_window:
object_area = numpy.sum(segmented_out > 0)
workspace.display_data.object_pct = (
100 * object_area / numpy.product(segmented_out.shape))
workspace.display_data.img = img
workspace.display_data.segmented_out = segmented_out
workspace.display_data.primary_labels = objects.segmented
workspace.display_data.global_threshold = global_threshold
workspace.display_data.object_count = object_count
def run(self, workspace):
"""Run the algorithm on one image set"""
#
# Get the image as a binary image
#
image_set = workspace.image_set
image = image_set.get_image(self.image_name.value, must_be_binary=True)
mask = image.pixel_data
if image.has_mask:
mask = mask & image.mask
angle_count = self.angle_count.value
#
# We collect the i,j and angle of pairs of points that
# are 3-d adjacent after erosion.
#
# i - the i coordinate of each point found after erosion
# j - the j coordinate of each point found after erosion
# a - the angle of the structuring element for each point found
#
i = numpy.zeros(0, int)
j = numpy.zeros(0, int)
a = numpy.zeros(0, int)
ig, jg = numpy.mgrid[0 : mask.shape[0], 0 : mask.shape[1]]
this_idx = 0
for angle_number in range(angle_count):
angle = float(angle_number) * numpy.pi / float(angle_count)
strel = self.get_diamond(angle)
erosion = binary_erosion(mask, strel)
#
# Accumulate the count, i, j and angle for all foreground points
# in the erosion
#
this_count = numpy.sum(erosion)
i = numpy.hstack((i, ig[erosion]))
j = numpy.hstack((j, jg[erosion]))
a = numpy.hstack((a, numpy.ones(this_count, float) * angle))
#
# Find connections based on distances, not adjacency
#
first, second = self.find_adjacent_by_distance(i, j, a)
#
# Do all connected components.
#
if len(first) > 0:
ij_labels = all_connected_components(first, second) + 1
nlabels = numpy.max(ij_labels)
label_indexes = numpy.arange(1, nlabels + 1)
#
# Compute the measurements
#
center_x = fixup_scipy_ndimage_result(
mean_of_labels(j, ij_labels, label_indexes)
)
center_y = fixup_scipy_ndimage_result(
mean_of_labels(i, ij_labels, label_indexes)
)
#
# The angles are wierdly complicated because of the wrap-around.
# You can imagine some horrible cases, like a circular patch of
# "worm" in which all angles are represented or a gentle "U"
# curve.
#
# For now, I'm going to use the following heuristic:
#
# Compute two different "angles". The angles of one go
# from 0 to 180 and the angles of the other go from -90 to 90.
# Take the variance of these from the mean and
# choose the representation with the lowest variance.
#
# An alternative would be to compute the variance at each possible
# dividing point. Another alternative would be to actually trace through
# the connected components - both overkill for such an inconsequential
# measurement I hope.
#
angles = fixup_scipy_ndimage_result(
mean_of_labels(a, ij_labels, label_indexes)
)
vangles = fixup_scipy_ndimage_result(
mean_of_labels(
(a - angles[ij_labels - 1]) ** 2, ij_labels, label_indexes
)
)
aa = a.copy()
aa[a > numpy.pi / 2] -= numpy.pi
aangles = fixup_scipy_ndimage_result(
mean_of_labels(aa, ij_labels, label_indexes)
)
vaangles = fixup_scipy_ndimage_result(
mean_of_labels(
(aa - aangles[ij_labels - 1]) ** 2, ij_labels, label_indexes
)
)
aangles[aangles < 0] += numpy.pi
angles[vaangles < vangles] = aangles[vaangles < vangles]
#
# Squish the labels to 2-d. The labels for overlaps are arbitrary.
#
labels = numpy.zeros(mask.shape, int)
labels[i, j] = ij_labels
else:
center_x = numpy.zeros(0, int)
center_y = numpy.zeros(0, int)
angles = numpy.zeros(0)
nlabels = 0
label_indexes = numpy.zeros(0, int)
labels = numpy.zeros(mask.shape, int)
m = workspace.measurements
assert isinstance(m, Measurements)
object_name = self.object_name.value
m.add_measurement(object_name, M_LOCATION_CENTER_X, center_x)
m.add_measurement(object_name, M_LOCATION_CENTER_Y, center_y)
m.add_measurement(object_name, M_ANGLE, angles * 180 / numpy.pi)
m.add_measurement(
object_name, M_NUMBER_OBJECT_NUMBER, label_indexes,
)
m.add_image_measurement(FF_COUNT % object_name, nlabels)
#
# Make the objects
#
object_set = workspace.object_set
assert isinstance(object_set, ObjectSet)
objects = Objects()
objects.segmented = labels
objects.parent_image = image
object_set.add_objects(objects, object_name)
if self.show_window:
workspace.display_data.i = center_y
workspace.display_data.j = center_x
workspace.display_data.angle = angles
workspace.display_data.mask = mask
workspace.display_data.labels = labels
workspace.display_data.count = nlabels
def test_shape(self):
objects = Objects()
objects.segmented = numpy.ones((224, 224), numpy.uint8)
assert objects.shape == (224, 224)
