def GenerateEdges(prefix, segmentation, subset, seg2gold_mapping=None):
# parameters from CVPR submission
network_radius = 600
maximum_distance = 500
width = (18, 52, 52)
# create the directory structure to save the features in
# forward is needed for training and validation data that is cropped
CreateDirectoryStructure(width, network_radius,
['training', 'validation', 'testing', 'forward'],
'edges')
# get the size of the data
zres, yres, xres = segmentation.shape
# make sure the subset is one of three categories
assert (subset == 'training' or subset == 'validation'
or subset == 'testing')
# crop the subset if it overlaps with testing data
((cropped_zmin, cropped_zmax), (cropped_ymin, cropped_ymax),
(cropped_xmin, cropped_xmax)) = dataIO.CroppingBox(prefix)
# call the function to actually generate the edges
edges = EndpointTraversal(prefix, segmentation, maximum_distance)
# create list for all relevant examples
positive_examples = []
negative_examples = []
unknown_examples = []
forward_positive_examples = []
forward_negative_examples = []
forward_unknown_examples = []
for edge in edges:
zpoint, ypoint, xpoint = (edge[IB_Z], edge[IB_Y], edge[IB_X])
label_one, label_two = edge[3], edge[4]
# if the center of the point falls outside the cropped box do not include it in training or validation
forward = False
# however, you allow it for forward inference
if (zpoint < cropped_zmin or cropped_zmax <= zpoint): forward = True
if (ypoint < cropped_ymin or cropped_ymax <= ypoint): forward = True
if (xpoint < cropped_xmin or cropped_xmax <= xpoint): forward = True
# see if these two segments belong to the same neuron
if not seg2gold_mapping is None:
gold_one = seg2gold_mapping[label_one]
gold_two = seg2gold_mapping[label_two]
else:
gold_one = -1
gold_two = -1
# create lists of locations where these point occur
if forward:
if gold_one < 1 or gold_two < 1:
forward_unknown_examples.append(edge)
elif gold_one == gold_two:
forward_positive_examples.append(edge)
else:
forward_negative_examples.append(edge)
else:
if gold_one < 1 or gold_two < 1:
unknown_examples.append(edge)
elif gold_one == gold_two:
positive_examples.append(edge)
else:
negative_examples.append(edge)
print 'No. Positive Edges: {}'.format(len(positive_examples))
print 'No. Negative Edges: {}'.format(len(negative_examples))
print 'No. Unknown Edges: {}'.format(len(unknown_examples))
parent_directory = 'features/biological/edges-{}nm-{}x{}x{}'.format(
network_radius, width[IB_Z], width[IB_Y], width[IB_X])
if len(positive_examples):
# save the examples
positive_filename = '{}/{}/positives/{}.examples'.format(
parent_directory, subset, prefix)
with open(positive_filename, 'wb') as fd:
fd.write(struct.pack('q', len(positive_examples)))
for ie, example in enumerate(positive_examples):
fd.write(
struct.pack('qqqqqq', example[0], example[1], example[2],
example[3], example[4], example[5]))
# create new examples array to remove last element
examples = []
for example in positive_examples:
examples.append(example[0:5])
positive_examples_array = GenerateExamplesArray(
prefix, segmentation, examples, width, network_radius)
dataIO.WriteH5File(positive_examples_array,
'{}/{}/positives/{}-examples.h5'.format(
parent_directory, subset, prefix),
'main',
compression=True)
del positive_examples_array
if len(negative_examples):
# save the examples
negative_filename = '{}/{}/negatives/{}.examples'.format(
parent_directory, subset, prefix)
with open(negative_filename, 'wb') as fd:
fd.write(struct.pack('q', len(negative_examples)))
for example in negative_examples:
fd.write(
struct.pack('qqqqqq', example[0], example[1], example[2],
example[3], example[4], example[5]))
# create new examples array to remove last element
examples = []
for example in negative_examples:
examples.append(example[0:5])
negative_examples_array = GenerateExamplesArray(
prefix, segmentation, examples, width, network_radius)
dataIO.WriteH5File(negative_examples_array,
'{}/{}/negatives/{}-examples.h5'.format(
parent_directory, subset, prefix),
'main',
compression=True)
del negative_examples_array
if len(unknown_examples):
# save the examples
unknown_filename = '{}/{}/unknowns/{}.examples'.format(
parent_directory, subset, prefix)
with open(unknown_filename, 'wb') as fd:
fd.write(struct.pack('q', len(unknown_examples)))
for example in unknown_examples:
fd.write(
struct.pack('qqqqqq', example[0], example[1], example[2],
example[3], example[4], example[5]))
# create new examples array to remove last element
examples = []
for example in unknown_examples:
examples.append(example[0:5])
unknown_examples_array = GenerateExamplesArray(prefix, segmentation,
examples, width,
network_radius)
dataIO.WriteH5File(unknown_examples_array,
'{}/{}/unknowns/{}-examples.h5'.format(
parent_directory, subset, prefix),
'main',
compression=True)
del unknown_examples_array
if len(forward_positive_examples):
# save the examples
forward_positive_filename = '{}/forward/positives/{}.examples'.format(
parent_directory, prefix)
with open(forward_positive_filename, 'wb') as fd:
fd.write(struct.pack('q', len(forward_positive_examples)))
for example in forward_positive_examples:
fd.write(
struct.pack('qqqqqq', example[0], example[1], example[2],
example[3], example[4], example[5]))
# create new examples array to remove last element
examples = []
for example in forward_positive_examples:
examples.append(example[0:5])
forward_positive_examples_array = GenerateExamplesArray(
prefix, segmentation, examples, width, network_radius)
dataIO.WriteH5File(forward_positive_examples_array,
'{}/forward/positives/{}-examples.h5'.format(
parent_directory, prefix),
'main',
compression=True)
del forward_positive_examples_array
if len(forward_negative_examples):
# save the examples
forward_negative_filename = '{}/forward/negatives/{}.examples'.format(
parent_directory, prefix)
with open(forward_negative_filename, 'wb') as fd:
fd.write(struct.pack('q', len(forward_negative_examples)))
for example in forward_negative_examples:
fd.write(
struct.pack('qqqqqq', example[0], example[1], example[2],
example[3], example[4], example[5]))
# create new examples array to remove last element
examples = []
for example in forward_negative_examples:
examples.append(example[0:5])
forward_negative_examples_array = GenerateExamplesArray(
prefix, segmentation, examples, width, network_radius)
dataIO.WriteH5File(forward_negative_examples_array,
'{}/forward/negatives/{}-examples.h5'.format(
parent_directory, prefix),
'main',
compression=True)
del forward_negative_examples_array
if len(forward_unknown_examples):
# save the examples
forward_unknown_filename = '{}/forward/unknowns/{}.examples'.format(
parent_directory, prefix)
with open(forward_unknown_filename, 'wb') as fd:
fd.write(struct.pack('q', len(forward_unknown_examples)))
for example in forward_unknown_examples:
fd.write(
struct.pack('qqqqqq', example[0], example[1], example[2],
example[3], example[4], example[5]))
# create new examples array to remove last element
examples = []
for example in forward_unknown_examples:
examples.append(example[0:5])
forward_unknown_examples_array = GenerateExamplesArray(
prefix, segmentation, examples, width, network_radius)
dataIO.WriteH5File(forward_unknown_examples_array,
'{}/forward/unknowns/{}-examples.h5'.format(
parent_directory, prefix),
'main',
compression=True)
del forward_unknown_examples_array
def CollapseGraph(prefix, segmentation, vertex_ones, vertex_twos,
maintained_edges, algorithm, evaluate):
# get the number of edges
nedges = maintained_edges.shape[0]
# create the union find data structure and collapse the graph
max_label = np.amax(segmentation) + 1
union_find = [unionfind.UnionFindElement(iv) for iv in range(max_label)]
# go through all of the edges
for ie in range(nedges):
# skip if the edge should not collapse
if maintained_edges[ie]: continue
# merge these vertices
vertex_one = vertex_ones[ie]
vertex_two = vertex_twos[ie]
unionfind.Union(union_find[vertex_one], union_find[vertex_two])
# create the mapping and save the result
mapping = np.zeros(max_label, dtype=np.int64)
for iv in range(max_label):
mapping[iv] = unionfind.Find(union_find[iv]).label
# apply the mapping and save the result
seg2seg.MapLabels(segmentation, mapping)
segmentation_filename = 'segmentations/{}-{}.h5'.format(prefix, algorithm)
dataIO.WriteH5File(segmentation, segmentation_filename, 'main')
# spawn a new meta file
dataIO.SpawnMetaFile(prefix, segmentation_filename, 'main')
# evaluate if gold data exists
if evaluate:
# get the variation of information for this result
new_prefix = segmentation_filename.split('/')[1][:-3]
# read in the new gold data
gold = dataIO.ReadGoldData(prefix)
rand_error, vi = comparestacks.VariationOfInformation(
new_prefix, segmentation, gold)
print 'Rand Error Full: {}'.format(rand_error[0] + rand_error[1])
print 'Rand Error Merge: {}'.format(rand_error[0])
print 'Rand Error Split: {}'.format(rand_error[1])
print 'Variation of Information Full: {}'.format(vi[0] + vi[1])
print 'Variation of Information Merge: {}'.format(vi[0])
print 'Variation of Information Split: {}'.format(vi[1])
if not os.path.exists('results'): os.mkdir('results')
with open('results/{}-{}.txt'.format(prefix, algorithm), 'w') as fd:
fd.write('Rand Error Full: {}\n'.format(rand_error[0] +
rand_error[1]))
fd.write('Rand Error Merge: {}\n'.format(rand_error[0]))
fd.write('Rand Error Split: {}\n'.format(rand_error[1]))
fd.write('Variation of Information Full: {}\n'.format(vi[0] +
vi[1]))
fd.write('Variation of Information Merge: {}\n'.format(vi[0]))
fd.write('Variation of Information Split: {}\n'.format(vi[1]))
def MergeGroundTruth(prefix, model_prefix):
# read the segmentation data
segmentation = dataIO.ReadSegmentationData(prefix)
# get the multicut filename (with graph weights)
multicut_filename = 'multicut/{}-{}.graph'.format(model_prefix, prefix)
# read the gold data
gold = dataIO.ReadGoldData(prefix)
# read in the segmentation to gold mapping
mapping = seg2gold.Mapping(segmentation, gold)
# get the maximum segmentation value
max_value = np.amax(segmentation) + 1
# create union find data structure
union_find = [UnionFind.UnionFindElement(iv) for iv in range(max_value)]
# read in all of the labels
with open(multicut_filename, 'rb') as fd:
# read the number of vertices and edges
nvertices, nedges, = struct.unpack('QQ', fd.read(16))
# read in all of the edges
for ie in range(nedges):
# read in the two labels
label_one, label_two, = struct.unpack('QQ', fd.read(16))
# skip over the reduced labels and edge weight
fd.read(24)
# if the labels are the same and the mapping is non zero
if mapping[label_one] == mapping[label_two] and mapping[label_one]:
UnionFind.Union(union_find[label_one], union_find[label_two])
# create a mapping
mapping = np.zeros(max_value, dtype=np.int64)
# update the segmentation
for iv in range(max_value):
label = UnionFind.Find(union_find[iv]).label
mapping[iv] = label
merged_segmentation = seg2seg.MapLabels(segmentation, mapping)
gold_filename = 'gold/{}_gold.h5'.format(prefix)
# TODO fix this code temporary filename
truth_filename = 'multicut/{}-truth.h5'.format(prefix)
# temporary write h5 file
dataIO.WriteH5File(merged_segmentation, truth_filename, 'stack')
import time
start_time = time.time()
print ('Ground truth: ')
# create the command line
command = '~/software/PixelPred2Seg/comparestacks --stack1 {} --stackbase {} --dilate1 1 --dilatebase 1 --relabel1 --relabelbase --filtersize 100 --anisotropic'.format(truth_filename, gold_filename)
# execute the command
os.system(command)
print (time.time() - start_time)
def Agglomerate(prefix, model_prefix, threshold=0.5):
# read the segmentation data
segmentation = dataIO.ReadSegmentationData(prefix)
# get the multicut filename (with graph weights)
multicut_filename = 'multicut/{}-{}.graph'.format(model_prefix, prefix)
# get the maximum segmentation value
max_value = np.amax(segmentation) + 1
# create union find data structure
union_find = [UnionFind.UnionFindElement(iv) for iv in range(max_value)]
# read in all of the labels and merge the result
with open(multicut_filename, 'rb') as fd:
# read the number of vertices and edges
nvertices, nedges, = struct.unpack('QQ', fd.read(16))
# read in all of the edges
for ie in range(nedges):
# read in both labels
label_one, label_two, = struct.unpack('QQ', fd.read(16))
# skip over the reduced labels
fd.read(16)
# read in the edge weight
edge_weight, = struct.unpack('d', fd.read(8))
# merge label one and label two in the union find data structure
if (edge_weight > threshold):
UnionFind.Union(union_find[label_one], union_find[label_two])
# create a mapping
mapping = np.zeros(max_value, dtype=np.int64)
# update the segmentation
for iv in range(max_value):
label = UnionFind.Find(union_find[iv]).label
mapping[iv] = label
# update the labels
agglomerated_segmentation = seg2seg.MapLabels(segmentation, mapping)
gold_filename = 'gold/{}_gold.h5'.format(prefix)
# TODO fix this code temporary filename
agglomeration_filename = 'multicut/{}-agglomerate.h5'.format(prefix)
# temporary - write h5 file
dataIO.WriteH5File(agglomerated_segmentation, agglomeration_filename, 'stack')
import time
start_time = time.time()
print ('Agglomeration - {}:'.format(threshold))
# create the command line
command = '~/software/PixelPred2Seg/comparestacks --stack1 {} --stackbase {} --dilate1 1 --dilatebase 1 --relabel1 --relabelbase --filtersize 100 --anisotropic'.format(agglomeration_filename, gold_filename)
# execute the command
os.system(command)
print (time.time() - start_time)
def Forward(prefix,
model_prefix,
segmentation,
subset,
seg2gold_mapping=None,
evaluate=False):
# optimal parameters from paper (width starts with 3 channels)
width = (3, 20, 60, 60)
radius = 400
threshold_volume = 10368000
# read in the trained model
model = model_from_json(open('{}.json'.format(model_prefix), 'r').read())
model.load_weights('{}-best-loss.h5'.format(model_prefix))
# get all of the examples
examples, npositives, nnegatives = CollectExamples(prefix, width, radius,
subset)
# get all of the large-small pairings
pairings = CollectLargeSmallPairs(prefix, width, radius, subset)
#assert (len(pairings) == examples.shape[0])
# get the threshold in terms of number of voxels
resolution = dataIO.Resolution(prefix)
threshold = int(threshold_volume /
(resolution[IB_Z] * resolution[IB_Y] * resolution[IB_X]))
# get the list of nodes over and under the threshold
small_segments, large_segments = FindSmallSegments(segmentation, threshold)
# get all of the probabilities
probabilities = model.predict_generator(NodeGenerator(examples, width),
examples.shape[0],
max_q_size=1000)
# save the probabilities to a file
output_filename = '{}-{}.probabilities'.format(model_prefix, prefix)
with open(output_filename, 'wb') as fd:
fd.write(struct.pack('q', examples.shape[0]))
for ie, (label_one, label_two) in enumerate(pairings):
fd.write(
struct.pack('qqd', label_one, label_two, probabilities[ie]))
# create the correct labels for the ground truth
ground_truth = np.zeros(npositives + nnegatives, dtype=np.bool)
for iv in range(npositives):
ground_truth[iv] = True
# get the results with labeled data
predictions = Prob2Pred(np.squeeze(probabilities[:npositives +
nnegatives]))
# print the confusion matrix
if not seg2gold_mapping is None:
output_filename = '{}-{}-inference.txt'.format(model_prefix, prefix)
PrecisionAndRecall(ground_truth, predictions, output_filename)
# create a mapping
small_segment_predictions = dict()
for small_segment in small_segments:
small_segment_predictions[small_segment] = set()
# go through each pairing
for pairing, probability in zip(pairings, probabilities):
label_one, label_two = pairing
# make sure that either label one or two is small and the other is large
assert ((label_one in small_segments) ^ (label_two in small_segments))
if label_one in small_segments:
small_segment = label_one
large_segment = label_two
else:
small_segment = label_two
large_segment = label_one
small_segment_predictions[small_segment].add(
(large_segment, probability[0]))
# begin to map the small labels
max_label = np.amax(segmentation) + 1
mapping = [iv for iv in range(max_label)]
if not seg2gold_mapping is None:
ncorrect_merges = 0
nincorrect_merges = 0
# go through all of the small segments
for small_segment in small_segments:
best_probability = -1
best_large_segment = -1
# go through all the neighboring large segments
for large_segment, probability in small_segment_predictions[
small_segment]:
if probability > best_probability:
best_probability = probability
best_large_segment = large_segment
# this should almost never happen but if it does just continue
if best_large_segment == -1 or best_probability < 0.5:
mapping[small_segment] = small_segment
continue
# get all of the best large segments
else:
mapping[small_segment] = best_large_segment
if not seg2gold_mapping is None:
# don't consider undetermined locations
if seg2gold_mapping[small_segment] < 1 or seg2gold_mapping[
best_large_segment] < 1:
continue
if seg2gold_mapping[small_segment] == seg2gold_mapping[
best_large_segment]:
ncorrect_merges += 1
else:
nincorrect_merges += 1
if not seg2gold_mapping is None:
print '\nResults:'
print ' Correctly Merged: {}'.format(ncorrect_merges)
print ' Incorrectly Merged: {}'.format(nincorrect_merges)
with open(output_filename, 'a') as fd:
fd.write('\nResults:\n')
fd.write(' Correctly Merged: {}\n'.format(ncorrect_merges))
fd.write(' Incorrectly Merged: {}\n'.format(nincorrect_merges))
# save the node mapping in the cache for later
end2end_mapping = [mapping[iv] for iv in range(max_label)]
# initiate the mapping to eliminate small segments
seg2seg.MapLabels(segmentation, mapping)
# reduce the labels and map again
mapping, _ = seg2seg.ReduceLabels(segmentation)
seg2seg.MapLabels(segmentation, mapping)
# update the end to end mapping with the reduced labels
for iv in range(max_label):
end2end_mapping[iv] = mapping[end2end_mapping[iv]]
# get the model name (first component is architecture and third is node-)
model_name = model_prefix.split('/')[1]
segmentation_filename = 'segmentations/{}-reduced-{}.h5'.format(
prefix, model_name)
dataIO.WriteH5File(segmentation, segmentation_filename, 'main')
# spawn a new meta file
dataIO.SpawnMetaFile(prefix, segmentation_filename, 'main')
# save the end to end mapping in the cache
if not os.path.exists('cache'): os.mkdir('cache')
mapping_filename = 'cache/{}-reduced-{}-end2end.map'.format(
prefix, model_name)
with open(mapping_filename, 'wb') as fd:
fd.write(struct.pack('q', max_label))
for label in range(max_label):
fd.write(struct.pack('q', end2end_mapping[label]))
if evaluate:
gold = dataIO.ReadGoldData(prefix)
new_prefix = segmentation_filename.split('/')[-1].split('.')[0]
# run the evaluation framework
rand_error, vi = comparestacks.VariationOfInformation(
new_prefix, segmentation, gold)
# write the output file
if not os.path.exists('results'): os.mkdir('results')
with open('results/{}-reduced-{}.txt'.format(prefix, model_name),
'w') as fd:
fd.write('Rand Error Full: {}\n'.format(rand_error[0] +
rand_error[1]))
fd.write('Rand Error Merge: {}\n'.format(rand_error[0]))
fd.write('Rand Error Split: {}\n'.format(rand_error[1]))
fd.write('Variation of Information Full: {}\n'.format(vi[0] +
vi[1]))
fd.write('Variation of Information Merge: {}\n'.format(vi[0]))
fd.write('Variation of Information Split: {}\n'.format(vi[1]))
