def hierarchical_merging_of_region_boundary_rags_example():
#img = data.coffee()
edges = sobel(skimage.color.rgb2gray(img))
labels = slic(img, compactness=30, n_segments=400)
g = graph.rag_boundary(labels, edges)
graph.show_rag(labels, g, img)
plt.title('Initial RAG')
labels2 = graph.merge_hierarchical(labels,
g,
thresh=0.08,
rag_copy=False,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
graph.show_rag(labels, g, img)
plt.title('RAG after hierarchical merging')
plt.figure()
out = skimage.color.label2rgb(labels2, img, kind='avg')
plt.imshow(out)
plt.title('Final segmentation')
plt.show()
def _quantify_2d(cls, tile, img_seg, nz, feature_calculators, **kwargs):
feature_values = []
for z in range(nz):
# Calculate properties of masked+labeled cell components
cell_props = measure.regionprops(img_seg[z][CytometerBase.CELL_MASK_CHANNEL], cache=False)
nucleus_props = measure.regionprops(img_seg[z][CytometerBase.NUCLEUS_MASK_CHANNEL], cache=False)
if len(cell_props) != len(nucleus_props):
raise ValueError(
'Expecting cell and nucleus properties to have same length (nucleus props = {}, cell props = {})'
.format(len(nucleus_props), len(cell_props))
)
# Compute RAG for cells if necessary
graph = None
if kwargs.get('cell_graph'):
labels = img_seg[z][CytometerBase.CELL_MASK_CHANNEL]
# rag_boundary fails on all zero label matrices so default to empty graph if that is the case
# see: https://github.com/scikit-image/scikit-image/blob/master/skimage/future/graph/rag.py#L386
if np.count_nonzero(labels) > 0:
graph = label_graph.rag_boundary(labels, np.ones(labels.shape))
else:
graph = label_graph.RAG()
# Loop through each detected cell and compute features
for i in range(len(cell_props)):
props = ObjectProperties(cell=cell_props[i], nucleus=nucleus_props[i])
# Run each feature calculator and add results in order
feature_values.append([
v for fc in feature_calculators
for v in fc.get_feature_values(tile, img_seg, graph, props, z)
])
return feature_values
def show_segmen_rag(array,
pathlibpath,
numSegments,
threshold,
cedges=177,
compactness=0.1,
sigma=5,
convert2lab=False):
print('Obtaining superstructures')
segments = slic(array,
compactness=compactness,
n_segments=numSegments,
sigma=sigma,
multichannel=False,
convert2lab=convert2lab)
print('Number of SV: ', len(np.unique(segments)))
segments += 1
edges = filters.sobel(array)
print('Obtaining RAG')
rag = graph.rag_boundary(segments, edges, connectivity=1)
segments2 = graph.merge_hierarchical(segments,
rag,
threshold=threshold,
in_place_merge=True,
rag_copy=False,
merge_func=merge_boundary,
weight_func=weight_boundary)
print('Final Number of SV: ', len(np.unique(segments2)))
graph.show_rag(segments, rag, array, edge_cmap='viridis')
save_yorno(array, segments, pathlibpath, cedges)
def get_patches(img, compactness=30, n_segments=200, rag_thresh=0.08):
"""Get list of patches from image found with SLIC."""
patches = []
img_lab = color.rgb2lab(img)
edges = filters.sobel(color.rgb2gray(img))
labels = segmentation.slic(img_lab, convert2lab=False,
compactness=compactness, n_segments=n_segments)
g = graph.rag_boundary(labels, edges)
segmented = graph.merge_hierarchical(labels, g, thresh=rag_thresh,
rag_copy=False,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
idxs_sorted = np.argsort(segmented.reshape((segmented.size)))
_, idxs_start = np.unique(
segmented.reshape((segmented.size))[idxs_sorted],
return_index=True)
idxs_for_values = np.split(idxs_sorted, idxs_start[1:])
for idxs in idxs_for_values:
rows, cols = np.unravel_index(idxs, img.shape[:2])
data = img[rows, cols]
lab = np.mean(img_lab[rows, cols], axis=0)
patches.append((rows, cols, data, lab))
return patches
def compute(self, image, n_region, mask=None):
"""
Parameters
----------
image: numpy ndarray
image array to be represented as regions
n_region : int
the number of regions to be generated
mask: mask image to give region of interest
Returns
-------
regions : a list of regions
"""
gray_image = image.copy()
if mask is not None:
gray_image[mask > 0] = 0
# Convert the original image to the 0~255 gray image
gray_image = (gray_image - gray_image.min()) / (gray_image.max() - gray_image.min())
labels = segmentation.slic(gray_image.astype(np.float),
n_segments=n_region,
slic_zero=True, sigma=2,
multichannel=False,
enforce_connectivity=True)
# edge_map = filters.sobel(gray_image) # just for 2-D image
edge_map = filters.laplace(gray_image)
# Given an image's initial segmentation and its edge map this method constructs the
# corresponding Region Adjacency Graph (RAG). Each node in the RAG represents a set of
# pixels within the image with the same label in labels. The weight between two adjacent
# regions is the average value in edge_map along their boundary.
rag = graph.rag_boundary(labels, edge_map)
regions = []
n_labels = labels.max() + 1
for r in np.arange(n_labels):
# get vox_pos
position = np.transpose(np.nonzero(labels == r))
# get vox_feat
vox_feat = np.zeros((position.shape[0], 1))
n_D = position.shape[1]
for i in range(position.shape[0]):
if n_D == 2:
vox_feat[i][0] = image[position[i][0], position[i][1]]
elif n_D == 3:
vox_feat[i][0] = image[position[i][0], position[i][1], position[i][2]]
else:
raise RuntimeError("We just consider 2_D and 3_D images at present!")
regions.append(Region(position, vox_feat=vox_feat, r_id=r))
for r in range(n_labels):
for r_key in rag.edge[r].keys():
regions[r].add_neighbor(regions[r_key])
self.regions = regions
self.image_shape = image.shape
return regions
def generate_neighbor_matrix(superpixel_map, sequence):
'''
This function is to convert the input superpixel image into a
matrix with shape (number_of_superpixel, number_of_superpixel).
It's used to indicate what's the neighbor of a superpixel.
"1" --> These two superpixels are neighbour.
"0" --> These two superpixels are not neightbour.
Arg:
superpixel_map: (Width, Height)
sequence: Segmentation values that corresponding to labels except background.
Return:
Adjacency matrix.
'''
pseudo_edge_map = np.ones(np.shape(superpixel_map))
rag = graph.rag_boundary(superpixel_map, pseudo_edge_map, connectivity=2)
num_super = rag.number_of_nodes()
neighbor_matrix = np.zeros((num_super, num_super))
nodes_list = rag.nodes()
for i in nodes_list:
if i not in sequence:
continue
for neighbor in rag.neighbors_iter(i):
neighbor_not_background = np.intersect1d(neighbor, sequence)
neighbor_matrix[i, neighbor_not_background] = 1
return neighbor_matrix
def get_groups(original):
"""
Finds a segmentation of image by taking an oversegmentation produced by the Priority-Flood watershed and
progressively reducing with a boundary region adjacency graph
:param original: Original RGB image to segment
:return: Segmented image. label = 0 represents an edge, label = -1 represents a pruned area
"""
original = gaussian(original, sigma=1.5, multichannel=True)
original = downscale_to(original, area_limit=2e5)
g = original[:, :, 1]
def weight_boundary(RAG, src, dst, n):
default = {'weight': 0.0, 'count': 0}
count_src = RAG[src].get(n, default)['count']
count_dst = RAG[dst].get(n, default)['count']
weight_src = RAG[src].get(n, default)['weight']
weight_dst = RAG[dst].get(n, default)['weight']
count = count_src + count_dst
return {
'count': count,
'weight': (count_src * weight_src + count_dst * weight_dst) / count
}
greyscale = rgb2gray(original)
gradient = np.hypot(sobel(greyscale, axis=0), sobel(greyscale, axis=1))
segmentation1 = watershed(gradient, markers=400, mask=greyscale > 0.3)
RAG = graph.rag_boundary(segmentation1, gradient)
segmentation2 = graph.merge_hierarchical(segmentation1,
RAG,
thresh=5e-3,
rag_copy=False,
in_place_merge=True,
merge_func=lambda *args: None,
weight_func=weight_boundary)
segmentation2[greyscale < 0.3] = -1
segmentation2 = prune(segmentation2,
abs_threshold=g.size / 1000,
default=-1)
counts, lo, hi = [], [], []
for label in set(segmentation2[segmentation2 >= 0]):
interior = distance_transform_edt(segmentation2 == label) >= 1.5
if np.sum(interior) >= 0:
counts.append(np.sum(interior))
lo.append(np.percentile(gradient[interior], q=70))
hi.append(np.percentile(gradient[interior], q=90))
edges = canny(greyscale,
low_threshold=np.average(lo, weights=counts),
high_threshold=np.average(hi, weights=counts))
edges = binary_dilation(edges, disk(2))
edges = binary_closing(edges, disk(5))
edges = remove_small_objects(edges, g.size / 1000)
edges = edges[1:-1, 1:-1]
edges = np.pad(edges, pad_width=1, mode='constant', constant_values=1)
groups = im_label(edges, background=1, connectivity=1)
groups = prune(groups, abs_threshold=g.size / 1000)
groups[greyscale <
0.15] = -2 # Ignore black areas due to mechanical vignetting
return groups
def boundaryRAG(cls, data, labels, sigma):
edges = sobel(data)
mat = np.copy(labels) + 1 if np.min(labels) == 0 else np.copy(labels)
g = graph.rag_boundary(mat, edges, connectivity=2)
props = regionprops(mat)
labelsToIndex = cls.labelsToPropsIndex(props)
cls.addData(g, data.shape, sigma, props, labelsToIndex)
return g
def process(self, data):
with Timer('superpixelnode_2'):
im_l, im_r, rgb, labels = data
edges = filters.sobel(color.rgb2gray(rgb))
g = graph.rag_boundary(labels, edges)
labels2 = graph.merge_hierarchical(labels,
g,
thresh=0.02,
rag_copy=False,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
return im_l, im_r, labels2.astype(np.uint8)
def get_segments(self, frame):
rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
labels = slic(rgb, self.n_segments)
edges = filters.sobel(color.rgb2gray(rgb))
g = graph.rag_boundary(labels, edges)
labels2 = graph.merge_hierarchical(labels,
g,
thresh=0.02,
rag_copy=False,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
return labels2.astype(np.uint8)
def merge_hier_boundary(labels, image, thresh=0.03, show_rag=False):
"""
Merges the given labels using a RAG based on boundaries.
Parameters
----------
labels: ndarray
image: ndarray
thresh: float
show_rag: bool
Returns
-------
rag: RAG
labels: ndarray
Merged labels.
"""
edges = filters.sobel(color.rgb2gray(image))
rag = graph.rag_boundary(labels, edges)
rag_copy = False
if show_rag:
rag_copy = True
fig, ax = plt.subplots(1, 2, figsize=(10, 10))
labels = graph.merge_hierarchical(labels,
rag,
thresh=thresh,
rag_copy=rag_copy,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
if show_rag:
graph.show_rag(labels, rag, image, ax=ax[0])
ax[0].title('Initial RAG')
graph.show_rag(labels, graph.rag_boundary(labels, edges), ax=ax[1])
ax[1].title('Final RAG')
return rag, labels
def region_boundry(image):
#image = imageGlobal
labels = segmentation.slic(image, compactness=30, n_segments=400)
edges = filters.sobel(image)
edges_rgb = color.gray2rgb(edges)
g = graph.rag_boundary(labels, edges)
lc = graph.show_rag(labels, g, edges_rgb, img_cmap=None, edge_cmap='viridis',
edge_width=1.2)
plt.colorbar(lc, fraction=0.03)
io.show()
def _get_neighbors_from_label_image(self,
suffix,
iterations=1,
save_neighbors=True,
**kwargs):
'''
From the image_label, find all_neighbors
Assumes that the labels on the image are the same as in the feature table.
On the label image, 0 is bg
objectNumbers start at 1
'''
labels = np.unique(
self.feature_table[self._column_objectnumber].values)
custom_image_label = self.image_label.copy()
custom_image_label[~np.isin(custom_image_label, labels)] = 0
if self._debug:
print(
"_get_neighbors_from_label_image\n#labels = {}; #objects = {}, starting label id = {}, iterations={}"
.format(len(labels), self.n, min(labels), iterations))
all_borders = (morphological_gradient(custom_image_label, size=3) > 0)
label_rag = graph.rag_boundary(custom_image_label,
all_borders.astype(float),
connectivity=iterations)
label_rag.remove_node(0)
adj_mat = np.array(nx.adjacency_matrix(label_rag).todense())
sum_neighbors = np.sum(adj_mat, axis=0)
self.adjacency_matrix[iterations] = [adj_mat]
if self._debug:
print("_get_neighbors_from_label_image", suffix)
print("_get_neighbors_from_label_image", len(sum_neighbors),
len(labels), self.n)
self.feature_table.loc[
self.feature_table[self._column_objectnumber].isin(labels),
'NumberNeighbors_{}'.format(suffix)] = sum_neighbors
if save_neighbors:
index_for_series = self.feature_table.index[self.feature_table[
self._column_objectnumber].isin(labels)]
self.feature_table.loc[:, 'neighbors_{}'.format(
suffix)] = self.feature_table[self._column_objectnumber].apply(
lambda v: label_rag.neighbors(v))
def test_rag_boundary():
labels = np.zeros((16, 16), dtype='uint8')
edge_map = np.zeros_like(labels, dtype=float)
edge_map[8, :] = 0.5
edge_map[:, 8] = 1.0
labels[:8, :8] = 1
labels[:8, 8:] = 2
labels[8:, :8] = 3
labels[8:, 8:] = 4
g = graph.rag_boundary(labels, edge_map, connectivity=1)
assert set(g.nodes()) == set([1, 2, 3, 4])
assert set(g.edges()) == set([(1, 2), (1, 3), (2, 4), (3, 4)])
assert g[1][3]['weight'] == 0.25
assert g[2][4]['weight'] == 0.34375
assert g[1][3]['count'] == 16
def test_rag_boundary():
labels = np.zeros((16, 16), dtype='uint8')
edge_map = np.zeros_like(labels, dtype=float)
edge_map[8, :] = 0.5
edge_map[:, 8] = 1.0
labels[:8, :8] = 1
labels[:8, 8:] = 2
labels[8:, :8] = 3
labels[8:, 8:] = 4
g = graph.rag_boundary(labels, edge_map, connectivity=1)
assert set(g.nodes()) == set([1, 2, 3, 4])
assert set(g.edges()) == set([(1, 2), (1, 3), (2, 4), (3, 4)])
assert g[1][3]['weight'] == 0.25
assert g[2][4]['weight'] == 0.34375
assert g[1][3]['count'] == 16
def get_rag(gray_image, labels):
"""
Get the region adjacency graph using the labeled image and corresponding the edge map generated by
greyscale image with sobel filter
:param gray_image: The greyscale image corresponding to the labeled image
:type gray_image: numpy.ndarray
:param labels: a labeled segmented image, where the pixels in the image are labeled with index of the \
segmented regions.
:type labels: numpy.ndarray
:return: The region adjacency graph in dictionary \
see https://scikit-image.org/docs/stable/auto_examples/segmentation/plot_rag_boundary.html for more \
references
:rtype: skimage.future.graph.RAG
"""
# Get rag of the labeled image
edge_map = sobel(gray_image)
rag = graph.rag_boundary(labels, edge_map)
return rag
def get_segments(self, frame):
if not self.initialized:
self.rgb = np.empty_like(frame)
self.lab = np.empty_like(frame)
self.initialized = True
#with Timer('super-labels1'):
cv2.cvtColor(frame, cv2.COLOR_BGR2RGB, dst=self.rgb)
labels = slic(self.rgb, self.n_segments)
#with Timer('super-labels2'):
edges = filters.sobel(color.rgb2gray(self.rgb))
g = graph.rag_boundary(labels, edges)
labels2 = graph.merge_hierarchical(labels,
g,
thresh=0.02,
rag_copy=False,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
return labels2.astype(np.uint8)
def rag_merge_threshold(edges, labels, threshold, size_pen):
'''
Merge adjacent segments using region adjacency graph based on strength of
edge between them.
'''
# create region adjacency graph using the edge info to produce
# edge weights between the nodes.
click.echo('creating Region Adjacency Graph')
rag = graph.rag_boundary(labels, edges)
# calculate pixel counts for each node
lab, counts = np.unique(labels.ravel(), return_counts=True)
click.echo('starting with {} segments'.format(lab.max()))
counts = (counts / counts.max()) * size_pen
for i, n in enumerate(lab):
rag.node[n].update({'pixels': counts[i]})
# update initial edge weights to weight by mean size of nodes
for n1, n2 in rag.edges_iter():
total_pix = rag.node[n1]['pixels'] + rag.node[n2]['pixels']
rag.edge[n1][n2]['weight'] += total_pix
rag.edge[n2][n1]['weight'] += total_pix
# calculate a value from the threshold percentile of edge weights.
edge_weights = [x[2]['weight'] for x in rag.edges(data=True)]
t = np.percentile(edge_weights, threshold)
# merge adjacent labels iteratively if their edge weight is below the
# required threshold.
click.echo('merging segments with edge weights below {} percentile'.format(
threshold))
refined_labels = graph.merge_hierarchical(
labels, rag, t, rag_copy=True, in_place_merge=True,
merge_func=merge_nodes, weight_func=update_edge_weights)
refined_labels, *_ = segmentation.relabel_sequential(refined_labels + 1,
offset=1)
click.echo('merged into {} segments'.format(refined_labels.max()))
return refined_labels
def region_boundry(img):
gimg = color.rgb2gray(img)
fig, ax = plt.subplots(nrows=1)
labels = segmentation.slic(img, compactness=30, n_segments=400)
edges = filters.sobel(gimg)
edges_rgb = color.gray2rgb(edges)
g = graph.rag_boundary(labels, edges)
lc = graph.show_rag(labels,
g,
edges_rgb,
img_cmap=None,
edge_cmap='viridis',
edge_width=1.2)
ax.imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
ax.set_title("Original Image")
plt.colorbar(lc, fraction=0.03)
plt.tight_layout()
plt.show()
def _ncutb_seg(self, data_list):
"""
This function use bounday instead of color to formulate the RAG
Use the ncut method to segment the image
[image, slic_label/[slic_param], [ncut_param]]
"""
img = data_list[0]
param = data_list[1]
param_cut = data_list[2]
if param_cut is None:
threahold = 0.2
else:
threahold = param_cut[0]
# Check if the param is the super pixel label or the num of super pixel
# to be segmented
try:
num = int(param[0])
# super pixel seg
label1 = segmentation.slic(img,
compactness=10,
n_segments=num,
max_iter=100,
slic_zero=True)
except:
label1 = param
# N-Cut
# Edge detection
edge = filters.sobel(color.rgb2gray(img))
# Smooth the edge map
edge = filters.gaussian(edge, 1)
edge = filters.gaussian(edge, 1)
# Reverse the energy map
ne = edge.max() - edge
rag = graph.rag_boundary(label1, ne)
label2 = graph.cut_normalized(label1, rag, thresh=threahold)
return label2
def make_network(mask, draw=False, save_loc=None):
orig, image_labels, edge_map = get_img_labels_edge_map(mask)
g = graph.rag_boundary(image_labels, edge_map)
g.remove_node(0)
if draw:
fig, ax = plt.subplots(2, 2, figsize=(15, 15), dpi=200)
ax = ax.ravel()
ax[0].imshow(orig, cmap='gray')
ax[2].imshow(label2rgb(image_labels, image=orig))
ax[1].imshow(edge_map)
lc = graph.show_rag(image_labels,
g,
edge_map,
ax=ax[3],
edge_width=5,
edge_cmap='Blues')
fig.colorbar(lc, fraction=0.03, ax=ax[3])
pos = {}
for idx in list(g.nodes):
pos[idx] = (np.array(g.nodes[idx]['centroid'])[::-1])
nx.draw(g, pos, ax=ax[3])
for a in ax:
a.grid('off')
fig.tight_layout()
if save_loc is not None:
fig.savefig(save_loc)
else:
# because if we don't draw then these features aren't added to the graph
props = regionprops(image_labels)
for (n, data), region, idx in zip(g.nodes(data=True), props,
range(len(props))):
data['centroid'] = tuple(map(int, region['centroid']))
data['uid'] = idx
return g
elif OVER_SEG == "quick":
segments = segmentation.quickshift(gray2rgb(img),
kernel_size=3,
max_dist=6,
ratio=0.5)
elif OVER_SEG == "water":
gradient = sobel(rgb2gray(img))
segments = segmentation.watershed(gradient, markers=400, compactness=0.001)
segments -= 1
else:
raise ValueError(OVER_SEG)
# Region Adjacency Graph
if ALGO == "hier":
edges = filters.sobel(color.rgb2gray(img))
g = graph.rag_boundary(segments, edges)
labels = graph.merge_hierarchical(segments,
g,
thresh=0.08,
rag_copy=True,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
elif ALGO == "ncut":
g = graph.rag_mean_color(img, segments, mode='similarity')
labels = graph.cut_normalized(segments,
g,
thresh=0.0002,
num_cuts=20,
in_place=False)
elif ALGO == "thresh":
def img_to_nodes(img, mask):
def weight_boundary(graph, src, dst, n):
"""
Handle merging of nodes of a region boundary region adjacency graph.
This function computes the `"weight"` and the count `"count"`
attributes of the edge between `n` and the node formed after
merging `src` and `dst`.
Parameters
----------
graph : RAG
The graph under consideration.
src, dst : int
The vertices in `graph` to be merged.
n : int
A neighbor of `src` or `dst` or both.
Returns
-------
data : dict
A dictionary with the "weight" and "count" attributes to be
assigned for the merged node.
"""
default = {'weight': 0.0, 'count': 0}
count_src = graph[src].get(n, default)['count']
count_dst = graph[dst].get(n, default)['count']
weight_src = graph[src].get(n, default)['weight']
weight_dst = graph[dst].get(n, default)['weight']
count = count_src + count_dst
return {
'count': count,
'weight': (count_src * weight_src + count_dst * weight_dst) / count
}
def merge_boundary(graph, src, dst):
"""Call back called before merging 2 nodes.
In this case we don't need to do any computation here.
"""
pass
def separate_regions(labels, graph):
indices = {}
for x in range(labels.shape[0]):
for y in range(labels.shape[1]):
id = labels[x][y]
if id not in indices:
indices[id] = []
indices[id].append((x, y))
nodes = {}
for key, value in indices.items():
nodes[key] = Node(key, graph, np.array(value))
for n in nodes.values():
n.update_neighbours(nodes, labels.size)
return nodes
def relabel(labels):
idx, counts = np.unique(labels, return_counts=True)
idx = idx[counts.argsort()]
idx = idx[::-1]
ch = np.zeros_like(idx)
ch[idx] = np.arange(idx.size)
labels = ch[labels]
return labels
edges = edg(img)
labels = me.label(mask)
g = graph.rag_boundary(labels, edges)
labels = graph.merge_hierarchical(labels,
g,
thresh=0.2,
rag_copy=False,
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
labels = relabel(labels)
labels = mp.dilation(labels)
labels = mp.area_closing(labels, 10)
labels = mp.erosion(labels)
g = graph.rag_boundary(labels, edges)
nodes = separate_regions(labels, g)
return labels, nodes
flat = np.nanargmin(voro_tensor, axis=0)
flat = np.ma.MaskedArray(flat, mask)
plt.figure(dpi=800)
plt.imshow(flat)
io.show()
edges = filters.sobel(flat)
edges_rgb = color.gray2rgb(edges)
from skimage.future import graph
plt.figure(dpi=800)
g = graph.rag_boundary(flat, edges, connectivity=1)
g.remove_node(0)
lc = graph.show_rag(flat, g, edges_rgb, img_cmap=None, edge_cmap='viridis',
edge_width=1.2)
plt.colorbar(lc, fraction=0.03)
plt.show()
for i in range(1, len(uni)):
print([uni[x] for x in g.neighbors(i)])
# plt.figure(dpi=500)
# plt.scatter(cols[rpos, 0], cols[rpos,1])
# plt.grid(True)
# plt.show()
def main(_):
if not FLAGS.dataset_dir:
raise ValueError('You must supply the dataset directory with --dataset_dir')
tf.logging.set_verbosity(tf.logging.INFO)
with tf.Graph().as_default():
tf_global_step = slim.get_or_create_global_step()
######################
# Gnerate the tfRecorder data #
######################
datareader,Volshape,rindex,spmap,labelOrg,imgOrg=EvalSample_Gnerate(FLAGS.dataset_dir, 'Glabel.nrrd')
Patchsize=len(rindex)
######################
# Select the dataset #
######################
dataset = dataset_factory.get_dataset(
FLAGS.dataset_name, FLAGS.dataset_split_name, FLAGS.dataset_dir,file_pattern=FLAGS.file_name,Datasize=Patchsize)
####################
# Select the model #
####################
#num_classes=2
with tf.Graph().as_default():
network_fn = nets_factory.get_network_fn(
FLAGS.model_name,
num_classes=(dataset.num_classes - FLAGS.labels_offset),
is_training=False)
##############################################################
# Create a dataset provider that loads data from the dataset #
##############################################################
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
shuffle=False,
common_queue_capacity=2 * FLAGS.batch_size,
common_queue_min=FLAGS.batch_size)
[image, label] = provider.get(['image', 'label'])
# label -= FLAGS.labels_offset
#####################################
# Select the preprocessing function #
#####################################
preprocessing_name = FLAGS.preprocessing_name or FLAGS.model_name
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name,
is_training=False)
eval_image_size = FLAGS.eval_image_size or network_fn.default_image_size
image = image_preprocessing_fn(image, eval_image_size, eval_image_size)
images, labels = tf.train.batch(
[image, label],
batch_size=FLAGS.batch_size,
num_threads=FLAGS.num_preprocessing_threads,
capacity=5 * FLAGS.batch_size)
###################
# Define the model #
####################
logits, end_points = network_fn(images)
probabilities = tf.nn.softmax(logits)
pred = tf.argmax(logits, dimension=1)
# if FLAGS.moving_average_decay:
# variable_averages = tf.train.ExponentialMovingAverage(
# FLAGS.moving_average_decay, tf_global_step)
# variables_to_restore = variable_averages.variables_to_restore(
# slim.get_model_variables())
# variables_to_restore[tf_global_step.op.name] = tf_global_step
# else:
# variables_to_restore = slim.get_variables_to_restore()
# #predictions = tf.argmax(logits, 1)
# labels = tf.squeeze(labels)
#
# # Define the metrics:
# names_to_values, names_to_updates = slim.metrics.aggregate_metric_map({
# 'Accuracy': slim.metrics.streaming_accuracy(predictions, labels),
# 'Recall@5': slim.metrics.streaming_recall_at_k(
# logits, labels, 5),
# })
#
# # Print the summaries to screen.
# summary_ops = []
# for name, value in names_to_values.items():
# summary_name = 'eval/%s' % name
# op = tf.scalar_summary(summary_name, value, collections=[])
# op = tf.Print(op, [value], summary_name)
# tf.add_to_collection(tf.GraphKeys.SUMMARIES, op)
# summary_ops.append(op)
# # TODO(sguada) use num_epochs=1
# if FLAGS.max_num_batches:
# num_batches = FLAGS.max_num_batches
# else:
# # This ensures that we make a single pass over all of the data.
# num_batches = math.ceil(dataset.num_samples / float(FLAGS.batch_size))
if tf.gfile.IsDirectory(FLAGS.checkpoint_path):
checkpoint_path = tf.train.latest_checkpoint(FLAGS.checkpoint_path)
else:
checkpoint_path = FLAGS.checkpoint_path
tf.logging.info('Evaluating %s' % checkpoint_path)
init_fn = slim.assign_from_checkpoint_fn(
os.path.join(checkpoint_path),
slim.get_model_variables())# vriable name???
#Volshape=(50,61,61)
imgshape=spmap.shape
segResult=np.zeros(imgshape)
groundtruth=np.zeros(imgshape)
segPromap=np.zeros(imgshape)
segPromapEdge=np.zeros(imgshape)
PreMap=[]
labellist=[]
seglist=[]
conv1list=[]
conv2list=[]
imgorglist=[]
fclist=[]
with tf.Session() as sess:
# Load weights
init_fn(sess)
# sess.run(images.initializer, feed_dict)
# Start input enqueue threads.
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
num_iter = int(math.ceil(dataset.num_samples/float(FLAGS.batch_size)))
step = 0
try:
while step < num_iter and not coord.should_stop():
# Run evaluation steps or whatever
segmentation,log,pmap,labelmap,imgpre,conv1,conv2,fc3= sess.run([pred,logits,probabilities,labels,images,end_points['conv1'],end_points['conv3'],end_points['fc3']])
step+=1
PreMap.extend(pmap)
conv1list.extend(conv1)
conv2list.extend(conv2)
fclist.extend(fc3)
imgorglist.extend(imgpre)
seglist.append(segmentation)
labellist.append(labelmap)
print('The No. %d/%d caculation'%(step,num_iter))
#Miaimshow.subplots(Pro_imgs, num=step+2, cols=8)
except tf.errors.OutOfRangeError:
print('Done evalutaion -- epoch limit reached')
finally:
# When done, ask the threads to stop.
coord.request_stop()
PreMap = np.array(PreMap)
np.save(os.path.join(FLAGS.dataset_dir,'liverspProbmap.npy'), PreMap)
# PreMap=np.squeeze(PreMap,axis=1)
# PreMap_flat = PreMap.ravel()
# PreMap_flat=np.divide((PreMap_flat - np.amin(PreMap_flat)) * 255, (np.amax(PreMap_flat) - np.amin(PreMap_flat)))
m = 0
for i in range(len(rindex)):
segResult[spmap==rindex[i]]=np.array(seglist).ravel()[i]
segPromap[spmap==rindex[i]]=PreMap[i,1]
segPromapEdge[spmap==rindex[i]]=PreMap[i,2]
groundtruth[spmap==rindex[i]]=np.array(labellist).ravel()[i]
coord.join(threads)
fig,ax= plt.subplots(nrows=2,ncols=3)
from skimage.segmentation import mark_boundaries
ax[0,0].set_title('Segmentation with superpixel map')
ax[0,0].imshow(mark_boundaries(segResult, spmap))
ax[0,1].set_title('Segmentation with ground truth map')
ax[0,1].imshow(segResult)
ax[0,1].imshow(labelOrg,alpha=0.5,cmap='jet')
ax[0,2].set_title('Reading label')
ax[0,2].imshow(groundtruth)
ax[1,0].set_title('liver Probabilities map')
ax[1,0].imshow(segPromap)
ax[1,1].set_title('edge Probabilities map')
ax[1,1].imshow(segPromapEdge)
ax[1,2].set_title('liver +edge Probabilities map')
ax[1,2].imshow(segPromapEdge+segPromap)
segthpro=segPromapEdge+segPromap
segthpro[segthpro<0.8]=0
from skimage.segmentation import active_contour
from skimage.measure import find_contours
from skimage.filters import gaussian
from skimage import morphology
#edg=sobel(segResult.astype(int))
segmorp=morphology.remove_small_objects(segthpro.astype(bool),5000)
segopen=morphology.opening(segmorp,morphology.disk(3))
segclose=morphology.closing(segopen,morphology.disk(15))
fig,ax=plt.subplots(1,3)
ax=ax.ravel()
ax[0].imshow(segmorp)
ax[0].set_title('Removed the small objects')
ax[1].imshow(segopen)
ax[1].set_title('After open operation')
ax[2].imshow(segclose)
ax[2].imshow(labelOrg,alpha=0.5,cmap='jet')
ax[2].set_title('After close operation')
plt.axis('off')
from MiaUtils import Miametrics as metric
mt=metric.MiaMetrics(logger)
dsc=mt.DSCMetric(segclose,labelOrg.astype(bool))
print('The dice similarity coefficient score is {}'.format(dsc))
voe=mt.VOEMetric(segclose,labelOrg.astype(bool))
print('The Volume overlap Error score is {}'.format(voe))
rvd=mt.RVDMetric(segclose,labelOrg.astype(bool))
print('The Relative voume difference score is {}'.format(rvd))
from medpy.metric.binary import hd
from medpy.metric.binary import asd
from medpy.metric.binary import obj_fpr
from medpy.metric.binary import obj_tpr
Asd=asd(segclose,labelOrg.astype(bool))
print('The Asd score is {}'.format(Asd))
HD=hd(segclose,labelOrg.astype(bool))
print('The Hausdorff Distance score is {}'.format(HD))
###************************************************************************
####superpixel-graph cuts mehtod computation
#######********************************************************************
from skimage import segmentation, color,filters
from skimage.future import graph
img=DataNormalize(imgOrg)/255
img=np.dstack((np.dstack((img, img)), img))
labels1 = segmentation.slic(img, compactness=5, n_segments=2000,sigma=1)
#labels1=spmap
out1 = color.label2rgb(labels1, img, kind='avg')
edge_map = filters.sobel(color.rgb2gray(img))
g = graph.rag_boundary(labels1, edge_map)
#g = graph.rag_mean_color(img, labels1, mode='similarity')
labels2 = graph.cut_normalized(labels1, g)
out2 = color.label2rgb(labels2, img, kind='avg')
fig, ax = plt.subplots(nrows=2, sharex=True, sharey=True)
ax[0].imshow(out1)
ax[1].imshow(out2)
for a in ax:
a.axis('off')
plt.tight_layout()
#array = np.transpose(array, (1, 0, 2))
print(array.shape)
print('Obtaining superstructures')
segments = slic(array,
compactness=compactness,
n_segments=numSegments,
multichannel=False,
convert2lab=convert2lab)
print('Number of SV: ', len(np.unique(segments)))
segments += 1
array = array.astype('int64')
mag = ndi.generic_gradient_magnitude(array, ndi.sobel, float)
mag *= 255.0 / np.max(mag) # normalize (Q&D)
print('Obtaining RAG')
rag = graph.rag_boundary(segments, mag, connectivity=1)
print('Merging RAG´s segments')
segments2 = graph.merge_hierarchical(segments,
rag,
253,
in_place_merge=True,
rag_copy=False,
merge_func=ji.merge_boundary,
weight_func=ji.weight_boundary)
print('Final Number of SV: ', len(np.unique(segments2)))
# 110
#graph.show_rag(segments, rag, array, edge_cmap= 'viridis')
destino_guardar = Path(
'C:/Users/Juan Ignacio/Documents/Movistar Cloud/TFM/Prueba_Feima_multiframe'
)
Construct a region boundary RAG with the ``rag_boundary`` function. The
function :py:func:`skimage.future.graph.rag_boundary` takes an
``edge_map`` argument, which gives the significance of a feature (such as
edges) being present at each pixel. In a region boundary RAG, the edge weight
between two regions is the average value of the corresponding pixels in
``edge_map`` along their shared boundary.
"""
from skimage.future import graph
from skimage import data, segmentation, color, filters, io
from matplotlib import pyplot as plt
img = data.coffee()
gimg = color.rgb2gray(img)
labels = segmentation.slic(img, compactness=30, n_segments=400, start_label=1)
edges = filters.sobel(gimg)
edges_rgb = color.gray2rgb(edges)
g = graph.rag_boundary(labels, edges)
lc = graph.show_rag(labels,
g,
edges_rgb,
img_cmap=None,
edge_cmap='viridis',
edge_width=1.2)
plt.colorbar(lc, fraction=0.03)
io.show()
)
imagen1 = "prueba_slic_100_pequena.tif"
imagen2 = 'prueba_slic_100.tif'
ruta1 = directorio / imagen2
prueba1 = ji.read_tiff(ruta1)
numSegments = 2200000 ## 1/30 #19000 prueba2d #100000 slic100 peque
#superp = slic(prueba1, n_segments = numSegments,compactness= 0.1, multichannel= False, convert2lab= False)
superp += 1
print('Number of SV: ', len(np.unique(superp)))
prueba1 = prueba1.astype('int64') # Precision
mag = ndi.generic_gradient_magnitude(prueba1, ndi.sobel,
float) # Float for rag_boundary
#mag *= 255.0 / np.max(mag) # normalize (Q&D)
ragb = graph.rag_boundary(superp, mag, connectivity=1)
# CLUST "CORTAR"
umbralc = 1000
superp_co = graph.cut_threshold(superp, ragb, umbralc)
print('Intermediate Number of SV: ', len(np.unique(superp_co)))
superp_co += 1
# CLUST "UNIR"
umbralb = 3500
ragb2 = graph.rag_boundary(superp_co, mag, connectivity=1)
superp_un = graph.merge_hierarchical(superp_co,
ragb2,
umbralb,
in_place_merge=True,
rag_copy=False,
merge_func=ji.merge_boundary,
weight_func=ji.weight_boundary)
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Fri Apr 7 10:13:38 2017
@author: luiz
"""
from skimage.future import graph
from skimage import data, segmentation, color, filters, io
from matplotlib import pyplot as plt
img = data.coffee()
gimg = color.rgb2gray(img)
labels = segmentation.slic(img, compactness=30, n_segments=400)
edges = filters.sobel(gimg)
edges_rgb = color.gray2rgb(edges)
g = graph.rag_boundary(labels, edges)
lc = graph.show_rag(labels, g, edges_rgb, img_cmap=None, edge_cmap='viridis',
edge_width=1.2)
plt.colorbar(lc, fraction=0.03)
io.show()
def process(path, N_SEGM, THRESH, ADAP):
im = Image.open(path) #'/home/olusiak/Obrazy/rois/41136_001.png-2.jpg')
#ran=1
#im_arr2 = np.fromstring(im.tobytes(), dtype=np.uint8)
#im_arr2 = im_arr2.reshape((im.size[1], im.size[0], 3))
#for x in range(len(im_arr2)):
# for y in range(len(im_arr2[0])):
# im_arr2[x][y][2]=0#=(im_arr2[0],255,255)
# im_arr2[x][y][1]=0
#plt.imshow(im_arr2)
#plt.show()
im_arr = np.fromstring(im.tobytes(), dtype=np.uint8)
im_arr = im_arr.reshape((im.size[1], im.size[0], 3))
gray = cv2.cvtColor(im_arr, cv2.COLOR_BGR2GRAY)
if ADAP:
thresh = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C,
cv2.THRESH_BINARY_INV, 55, 20)
else:
data = dict()
for i in gray.ravel():
if data.has_key(i):
data[i] += 1
else:
data[i] = 1
# if data.has_key(255):
max = 0
id = None
for k in data.keys():
# print 'k ',k
if data[k] >= max and k < 240 and k > 130: # 50:#240:
id = k
max = data[k]
# for k in data.keys():
# print 'k ',k,' ',data[k]
# print 'key: ',id,' - ',max
id -= THRESH # 35
ret, thresh = cv2.threshold(
gray, id, 255, cv2.ADAPTIVE_THRESH_MEAN_C + cv2.THRESH_BINARY_INV
) #+cv2.THRESH_OTSU)#cv2.ADAPTIVE_THRESH_MEAN_C+
#plt.matshow(thresh,cmap='gray')
#plt.show()
kernel = np.ones((3, 3), np.uint8)
opening = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=3)
opening2 = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=3)
#plt.imshow(opening)
#plt.show()
for i in range(opening.shape[0]):
for j in range(opening.shape[1]):
if opening[i][j] == 0:
im_arr[i][j] = (0, 0, 0)
#plt.imshow(im_arr)
#plt.show()
#plt.matshow(opening)
#plt.show()
#plt.imshow(im_arr)
#plt.show()
#im = Image.fromarray(im_arr)
#im.thumbnail((im.size[0] / ran, im.size[1] / ran), Image.ANTIALIAS)
#im_arr = np.fromstring(im.tobytes(), dtype=np.uint8)
#im_arr = im_arr.reshape((im.size[1], im.size[0], 3))
#im = Image.fromarray(opening)
#im.thumbnail((im.size[0] / ran, im.size[1] / ran), Image.ANTIALIAS)
#opening = np.fromstring(im.tobytes(), dtype=np.uint8)
#opening = opening.reshape((im.size[1], im.size[0]))
img2 = im_arr
edges = filters.sobel(color.rgb2gray(img2))
labels = segmentation.slic(img2, compactness=30,
n_segments=N_SEGM) #30 2000
#g = graph.rag_mean_color(img2, labels)#added
g = graph.rag_boundary(labels, edges) #first
#graph.show_rag(labels, g, img)
#plt.title('Initial RAG')
labels2 = graph.merge_hierarchical(
labels,
g,
thresh=0.98,
rag_copy=False, # 0.08
in_place_merge=True,
merge_func=merge_boundary,
weight_func=weight_boundary)
#final_labels = graph.cut_threshold(labels, g, 29)#added
#final_label_rgb = color.label2rgb(final_labels, img2, colors=cols, kind='overlay')#added
#labels2=final_label_rgb#added
#plt.imshow(final_label_rgb)
#plt.show()
#graph.show_rag(labels, g, im)
#plt.title('RAG after hierarchical merging')
#plt.figure()
#ret, opening = cv2.threshold(opening,0,255,cv2.THRESH_OTSU)#cv2.ADAPTIVE_THRESH_MEAN_C+
#out = color.label2rgb(labels2, img2, kind='avg')
s = set()
for row in labels2:
for e in row:
s.add(e)
#print 'sss ',len(s)
cols = list()
c = 0
cp = len(s) + 5
for r in range(0, 256, 1):
for r2 in range(0, 256, 1):
for r3 in range(0, 256, 1):
cols.append((r, r2, r3))
cp -= 1
if cp == 0:
break
if cp == 0:
break
if cp == 0:
break
# print 'cols', len(cols)
shuffle(cols)
img2 = np.zeros_like(img2)
img2[:, :, 0] = opening2
img2[:, :, 1] = opening2
img2[:, :, 2] = opening2
out = color.label2rgb(labels2, img2, colors=cols, kind='overlay', alpha=1)
for i in range(img2.shape[0]):
for j in range(img2.shape[1]):
if img2[i][j][0] == 0 and img2[i][j][1] == 0 and img2[i][j][
2] == 0: #,0,0]:
out[i][j][0] = 0
out[i][j][1] = 0
out[i][j][2] = 0
#print 'OUT'
#plt.imshow(out)
#plt.show()
#plt.imshow(out)
#plt.show()
#xx = set()
#for i in range(out.shape[0]): # fx.shape[0]):
# for j in range(out.shape[1]): # fx.shape[1]):
# s = (out[i, j][0], out[i, j][1], out[i, j][2])
# xx.add(s)
out = np.uint8(out)
im = Image.fromarray(out)
#plt.imshow(out)
#plt.show()
return im
for l in files: s = "pic/Expressionism/" s += str(l) print(s) f.write(s + ' ') img = io.imread(s) img = cv.GaussianBlur(img, (11,11), 0) labels = segmentation.slic(img, compactness=30, n_segments=150) labels = labels + 1 regions = regionprops(labels) edge_map = filters.sobel(color.rgb2gray(img)) label_rgb = color.label2rgb(labels, img, kind='avg') label_rgb = segmentation.mark_boundaries(label_rgb, labels, (0, 1, 1)) edge_map = filters.sobel(color.rgb2gray(img)) rag = graph.rag_boundary(labels, edge_map) for region in regions: rag.node[region['label']]['centroid'] = region['centroid'] rag.node[region['label']]['color'] = label_rgb[region['centroid']] red = np.asarray([1, 0, 0]) green = np.asarray([0, 1, 0]) blue = np.asarray([0, 0, 1]) yellow = np.asarray([1, 1, 0]) purple = np.asarray([0.5, 0, 0.5]) orange = np.asarray([1, 0.5, 0]) palette = [red, green, blue, yellow, purple, orange] harmony = 0 disharmony = 0 tr = 0.4
def create_graph(floormap, return_dist=False, room_coordinates=False):
"""
Segment the floormap into rooms and create a Region Adjacency Graph for the level.
Many settings for decorating the graph are possible, by default the simplest form is returned.
:param floormap: Binary image representing the level floor
:param return_dist: If true, also returns the distance map of each point to the closest wall.
:param room_coordinates: If true, each graph node will contain the room vertices and information about the walls
:return: (Roommap, Graph) if return_dist is False, (Roommap, Graph, dist) otherwise
"""
# Ensuring that floormap is always a boolean array
floormap = floormap.astype(np.bool)
#floormap = rescale(floormap, 2)
dist = ndi.distance_transform_edt(floormap)
threshold = int(dist.max())
optimal_threshold = 0
number_of_centers = 0
# Finding room center and finding the optimal threshold (the one that maximizes the number of rooms)
for i in range(int(dist.max()), int(dist.min()) - 1, -1):
local_max = peak_local_max(dist,
threshold_abs=threshold - i,
indices=False,
labels=floormap,
min_distance=3)
markers = ndi.label(local_max)[0]
if markers.max() > number_of_centers:
optimal_threshold = threshold - i
number_of_centers = markers.max()
# Computing roommap with the optimal threshold
local_max = peak_local_max(dist,
min_distance=3,
indices=False,
labels=floormap,
threshold_abs=optimal_threshold)
markers = ndi.label(local_max)[0]
roommap = watershed(-dist, markers, mask=floormap)
room_RAG_boundaries = skg.rag_boundary(
roommap, filters.sobel(color.rgb2gray(roommap)))
if room_coordinates:
# For each floor...
floors = label(floormap)
for floor_id in range(max(1, floors.min()),
floors.max() +
1): # Skipping label 0 (background)
# Building the wall list for floor boundaries
# Here the map is upsampled by a factor 2 before finding the contours, then coordinates are divided by two.
# This is for avoiding "X" shaped connections between rooms due to how find_contours work
floor_contour = find_contours(resize(
floors == floor_id, (floors.shape[0] * 2, floors.shape[1] * 2),
order=0),
0.5,
positive_orientation='low')[0] / 2
walls_vertices = [tuple(v) for v in floor_contour]
floor_boundaries = tuple(vertices_to_segment_list(walls_vertices))
# Map of rooms belonging to current floor
rooms = roommap * (floors == floor_id)
for room_id in range(max(1, rooms.min()),
rooms.max() +
1): # Skipping label 0 (background)
if room_id not in rooms:
# Some room id may be in another floor, if they are enumerated horizontally
continue
# Here the map is upsampled by a factor 2 before finding the contours, then coordinates are divided by two.
# This is for avoiding "X" shaped connections between rooms due to how find_contours work
room_contour = find_contours(resize(
rooms == room_id, (rooms.shape[0] * 2, rooms.shape[1] * 2),
order=0),
0.5,
fully_connected='high',
positive_orientation='low')[0] / 2
rooms_vertices = [tuple(v) for v in room_contour]
room_boundaries = tuple(
vertices_to_segment_list(rooms_vertices))
room_RAG_boundaries.node[room_id]['walls'] = list()
for segment in room_boundaries:
leads_to = 0 if segment in floor_boundaries else None # We cannot still know edges for other rooms but background
room_RAG_boundaries.node[room_id]['walls'].append(
(segment, leads_to))
# Here we still miss the relation between boundary and edges.
# Second pass
for room_id in range(max(1, rooms.min()), rooms.max() + 1):
if room_id not in rooms:
# Some room id may be in another floor, if they are enumerated horizontally
continue
boundaries_current = {
wall
for wall in room_RAG_boundaries.node[room_id]['walls']
if wall[1] is None
}
for neigh in room_RAG_boundaries.adj[room_id]:
if neigh == 0:
continue
# Finding the neighbour boundaries. We must consider both directions for each vertex
boundaries_neigh = {
wall
for wall in room_RAG_boundaries.node[neigh]['walls']
if wall[1] is None
}
boundaries_neigh_reverse = {
_reverse_wall(wall)
for wall in room_RAG_boundaries.node[neigh]['walls']
if wall[1] is None
}
common_segments = boundaries_current.intersection(
boundaries_neigh)
common_segments_reversed = boundaries_current.intersection(
boundaries_neigh_reverse)
# Marking the boundary in the two nodes with the destination node
# Each node will contain the list
for cs in common_segments:
i_current = room_RAG_boundaries.node[room_id][
'walls'].index(cs)
i_neighbour = room_RAG_boundaries.node[neigh][
'walls'].index(cs)
room_RAG_boundaries.node[room_id]['walls'][
i_current] = (cs[0], neigh)
room_RAG_boundaries.node[neigh]['walls'][
i_neighbour] = (cs[0], room_id)
# Same thing in the case of reversed segments
for cs in common_segments_reversed:
rev_cs = _reverse_wall(cs)
i_current = room_RAG_boundaries.node[room_id][
'walls'].index(cs)
i_neighbour = room_RAG_boundaries.node[neigh][
'walls'].index(rev_cs)
room_RAG_boundaries.node[room_id]['walls'][
i_current] = (cs[0], neigh)
room_RAG_boundaries.node[neigh]['walls'][
i_neighbour] = (rev_cs[0], room_id)
if return_dist:
return roommap, room_RAG_boundaries, dist
return roommap, room_RAG_boundaries
def quantify(self, tile, img_seg, channel_names=None,
include_cell_intensity=True,
include_nucleus_intensity=False,
include_cell_graph=False,
spot_count_channels=None,
spot_count_params=None):
ncyc, nz, _, nh, nw = tile.shape
# Move cycles and channels to last axes (in that order)
tile = np.moveaxis(tile, 0, -1)
tile = np.moveaxis(tile, 1, -1)
# Collapse tile to ZHWC (instead of cycles and channels being separate)
tile = np.reshape(tile, (nz, nh, nw, -1))
nch = tile.shape[-1]
# Generate default channel names list if necessary
if channel_names is None:
channel_names = ['{:03d}'.format(i) for i in range(nch)]
if nch != len(channel_names):
raise ValueError(
'Tile has {} channels but given channel name list has {} (they should be equal); '
'channel names given = {}, tile shape = {}'
.format(nch, len(channel_names), channel_names, tile.shape)
)
# Configure features to be calculated based on provided flags
feature_calculators = [BasicCellFeatures()]
if include_cell_intensity:
feature_calculators.append(IntensityFeatures(nch, channel_names, COMP_CELL))
if include_nucleus_intensity:
feature_calculators.append(IntensityFeatures(nch, channel_names, COMP_NUCLEUS))
if include_cell_graph:
feature_calculators.append(GraphFeatures())
if spot_count_channels is not None:
indexes = [channel_names.index(c) for c in spot_count_channels]
params = spot_count_params or {}
feature_calculators.append(SpotFeatures(indexes, spot_count_channels, **params))
# Compute list of resulting feature names (values will be added in this order)
feature_names = [v for fc in feature_calculators for v in fc.get_feature_names()]
feature_values = []
for z in range(nz):
# Calculate properties of masked+labeled cell components
cell_props = measure.regionprops(img_seg[z][CELL_CHANNEL], cache=False)
nucleus_props = measure.regionprops(img_seg[z][NUCLEUS_CHANNEL], cache=False)
if len(cell_props) != len(nucleus_props):
raise ValueError(
'Expecting cell and nucleus properties to have same length (nucleus props = {}, cell props = {})'
.format(len(nucleus_props), len(cell_props))
)
# Compute RAG for cells if necessary
graph = None
if include_cell_graph:
labels = img_seg[z][CELL_CHANNEL]
# rag_boundary fails on all zero label matrices so default to empty graph if that is the case
# see: https://github.com/scikit-image/scikit-image/blob/master/skimage/future/graph/rag.py#L386
if np.count_nonzero(labels) > 0:
graph = label_graph.rag_boundary(labels, np.ones(labels.shape))
else:
graph = label_graph.RAG()
# Loop through each detected cell and compute features
for i in range(len(cell_props)):
props = ObjectProperties(cell=cell_props[i], nucleus=nucleus_props[i])
# Run each feature calculator and add results in order
feature_values.append([
v for fc in feature_calculators
for v in fc.get_feature_values(tile, img_seg, graph, props, z)
])
return pd.DataFrame(feature_values, columns=feature_names)
if not in_place:
self.remove_node(dst)
return new
graph.RAG.merge_nodes = merge_nodes
# <markdowncell>
# Now we can make a RAG that will be mergeable:
# <codecell>
g = graph.rag_boundary(ws, edges)
# <markdowncell>
# g is now a *graph* in which each region is a node, and each node links to that
# regions neighbors. The edges have hold properties about the boundary between
# the corresponding region:
# <codecell>
plt.imshow(ws == 45)
print(g[45])