/
segmentation.py
551 lines (466 loc) · 16.6 KB
/
segmentation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
"""
Functions to perform segmentation of NMR spectrum.
"""
import numpy as np
import numpy.ma as ma
import scipy.ndimage as ndimage
from .analysisbase import neighbors
# Connected segmenting method:
# The connected segmentation method finds all nodes which are above a given
# threshold and connected to the initial point. For finding all segments
# the scipy.ndimage.label function is used for speed.
def label_connected(data, thres, structure):
"""
Label connected features in data. Returns labeled_array, num_features
"""
return ndimage.label(data > thres, structure)
def find_all_connected(data, thres, find_segs=False, diag=False):
"""
Find all connected segments.
Parameters
----------
data : ndarray
Data to perform segmentation on.
thres : float
Threshold, below this nodes are considered noise.
find_segs : bool, optional
True to return a list of slices for the segments.
diag : bool
True to include diagonal neighbors in connection.
Returns
-------
locations : list
List of indicies of local maximum in each segment.
seg_slices : list, optional
List of slices which extract a given segment from the data. Only
returned when fig_segs is True.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
# determine labeled array of segments
labels, num_features = label_connected(data, thres, structure)
# determine locations of segment maxima
locations = ndimage.maximum_position(data, labels, range(1, num_features +
1))
# find segment slices if requested and return
if find_segs == True:
seg_slices = ndimage.find_objects(labels)
return locations, seg_slices
else:
return locations
# nconnected method:
# The nconnected method is identical to the connected method except nodes must
# be below the threshold and local minimum are reported. This is useful for
# finding negative peaks by setting thres to the negative of the noise level.
def label_nconnected(data, thres, structure):
"""
Label nconnected features in data. Returns labeled_array, num_features
"""
return ndimage.label(data < thres, structure)
def find_all_nconnected(data, thres, find_segs=False, diag=False):
"""
Find all negatively connected segments in data.
Parameters
----------
data : ndarray
Data to perform segmentation on.
thres : float
Threshold, below this nodes are considered noise.
find_segs : bool, optional
True to return a list of slices for the segments.
diag : bool
True to include diagonal neighbors in connection.
Returns
-------
locations : list
List of indicies of local maximum in each segment.
seg_slices : list, optional
List of slices which extract a given segment from the data. Only
returned when fig_segs is True.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
# determine labeled array of segments
labels, num_features = label_nconnected(data, thres, structure)
# determine locations of segment maxima
locations = ndimage.minimum_position(data, labels, range(1,
num_features + 1))
# find segment slices if requested and return
if find_segs == True:
seg_slices = ndimage.find_objects(labels)
return locations, seg_slices
else:
return locations
# downward segmentation method:
# The downward segmenting method uses the flood fill algorithm to find
# all points connected to an initial node which are above a given threshold
# and to which a path exists in which each step of the path moves lower in
# intensity. This can be though of as all points accessible by a water drop
# following downward slopes from the initial node.
# Upward segmentation uses the same priciple except nodes must be below
# the threshold an upward path must exist.
def mark_dseg(mdata, map, pt, mark, structure):
"""
Mark downward-connected region on segment map starting at node pt.
Modifies mdata mask and map.
Parameters
----------
mdata : masked ndarray
Masked data array.
map :
Array mapping out segments.
pt : tuple of ints
Index of starting node
mark : int
Integer to mark map with.
"""
if mdata.mask[pt] == True:
return
else:
map[pt] = mark
mdata[pt] = ma.masked
Q = [pt]
while Q:
pt = Q.pop(0)
v = mdata.data[pt]
# Check all neightbors
for new_pt in neighbors(pt, mdata.shape, structure):
if mdata.mask[new_pt] == False and mdata[new_pt] < v:
Q.append(new_pt)
map[new_pt] = mark
mdata[new_pt] = ma.masked
return
def label_downward_seg(data, labels, seg_slice, seg_index, max_index,
structure):
""" Label a segment which is downward connected """
slabels = labels[seg_slice]
msdata = np.ma.masked_array(data[seg_slice], mask=(slabels != seg_index))
# mark the downward connected segment with the highest peak in the
# selected segment with the segment index.
argmax = np.unravel_index(msdata.argmax(), msdata.shape)
mark_dseg(msdata, slabels, argmax, seg_index, structure)
# mark any
while msdata.mask.all() == False:
argmax = np.unravel_index(msdata.argmax(), msdata.shape)
mark_dseg(msdata, slabels, argmax, max_index, structure)
max_index = max_index + 1
return max_index
def label_downward(data, thres, structure):
"""
Label connected features in data. Returns labeled_array, num_features
"""
# find connected segments
labels, num_features = ndimage.label(data > thres, structure)
seg_slices = ndimage.find_objects(labels)
max_index = int(num_features + 1)
# loop over the segments and perform a downward segment on each
for i, s in enumerate(seg_slices):
max_index = label_downward_seg(data, labels, s, i + 1, max_index,
structure)
return labels, max_index - 1
def find_all_downward(data, thres, find_segs=False, diag=False):
"""
Find all downward connected segments in data
Parameters
----------
data : ndarray
Data to perform segmentation on.
thres : float
Threshold, below this nodes are considered noise.
find_segs : bool, optional
True to return a list of slices for the segments.
diag : bool, optional
True to include diagonal neighbors in connection.
Returns
-------
locations : list
List of indicies of local maximum in each segment.
seg_slices : list, optional
List of slices which extract a given segment from the data. Only
returned when fig_segs is True.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
# determine labeled array of segments
labels, num_features = label_downward(data, thres, structure)
# determine locations of segment maxima
locations = ndimage.maximum_position(data, labels, range(1,
num_features + 1))
# find segment slices if requested and return
if find_segs == True:
seg_slices = ndimage.find_objects(labels)
return locations, seg_slices
else:
return locations
def mark_useg(mdata, map, pt, mark, structure):
"""
Mark upward-connected region on segment map starting at node pt
Modifies mdata mask and map.
Parameters
----------
mdata : masked ndarray
Masked data array.
map :
Array mapping out segments.
pt : tuple of ints
Index of starting node
mark : int
Integer to mark map with.
"""
if mdata.mask[pt] == True:
return
else:
map[pt] = mark
mdata[pt] = ma.masked
Q = [pt]
while Q:
pt = Q.pop(0)
v = mdata.data[pt]
# Check all neightbors
for new_pt in neighbors(pt, mdata.shape, structure):
if mdata.mask[new_pt] == False and mdata[new_pt] > v:
Q.append(new_pt)
map[new_pt] = mark
mdata[new_pt] = ma.masked
return
def label_upward_seg(data, labels, seg_slice, seg_index, max_index,
structure):
""" Label a segment which is upward connected """
slabels = labels[seg_slice]
msdata = np.ma.masked_array(data[seg_slice],
mask = (slabels != seg_index))
# mark the upward connected segment with the highest peak in the
# selected segment with the segment index.
argmin = np.unravel_index(msdata.argmin(), msdata.shape)
mark_useg(msdata, slabels, argmin, seg_index, structure)
# mark any
while msdata.mask.all() == False:
argmin = np.unravel_index(msdata.argmin(), msdata.shape)
mark_useg(msdata, slabels, argmin, max_index, structure)
max_index = max_index + 1
return max_index
def label_upward(data, thres, structure):
"""
Label upward connected features in data. Returns labeled_array,
num_features
"""
# find connected segments
labels, num_features = ndimage.label(data < thres, structure)
seg_slices = ndimage.find_objects(labels)
max_index = int(num_features + 1)
# loop over the segments and perform a downward segment on each
for i, s in enumerate(seg_slices):
max_index = label_upward_seg(data, labels, s, i + 1, max_index,
structure)
return labels, max_index - 1
def find_all_upward(data, thres, find_segs=False, diag=False):
"""
Find all upward connected segments in data
Parameters
----------
data : ndarray
Data to perform segmentation on.
thres : float
Threshold, below this nodes are considered noise.
find_segs : bool, optional
True to return a list of slices for the segments.
diag : bool, optional
True to include diagonal neighbors in connection.
Returns
-------
locations : list
List of indicies of local maximum in each segment.
seg_slices : list, optional
List of slices which extract a given segment from the data. Only
returned when fig_segs is True.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
# determine labeled array of segments
labels, num_features = label_upward(data, thres, structure)
# determine locations of segment maxima
locations = ndimage.minimum_position(data, labels,
range(1, num_features + 1))
# find segment slices if requested and return
if find_segs == True:
seg_slices = ndimage.find_objects(labels)
return locations, seg_slices
else:
return locations
##########################
# Single point functions #
##########################
def find_downward(data, pt, thres, diag=False):
"""
Find points downward-connected to a point in data.
Parameters
----------
data : ndarray
Array of data.
pt : tuple of ints
Starting point of peak.
thres : float
Threshold, below this nodes are considered noise.
diag : bool, optional
True to include diagonal neighbors in connection.
Returns
-------
nodes : list
Indicies of downward-connected nodes.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
if type(pt) == int:
pt = (pt, )
pt = tuple(pt)
shape = data.shape
if data[pt] < thres: # check that the initial point is above threshold.
return []
Q = [pt] # queue
segment = [pt]
while Q: # loop until Q is empty
pt = Q.pop(0) # remove first element of queue
v = data[pt] # value at current node
for new_pt in neighbors(pt, shape, structure): # check all neightbors
if thres < data[new_pt] < v and new_pt not in segment:
Q.append(new_pt)
segment.append(new_pt)
return segment
def find_connected(data, pt, thres, diag=False):
"""
Find points connected to a point in data.
Parameters
----------
data : ndarray
Array of data.
pt : tuple of ints
Starting point of peak.
thres : float
Threshold, below this nodes are considered noise.
diag : bool, optional
True to include diagonal neighbors in connection.
Returns
-------
nodes : list
Indicies of connected nodes.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
if type(pt) == int:
pt = (pt, )
pt = tuple(pt)
shape = data.shape
if data[pt] < thres: # check that the initial point is above threshold.
return []
Q = [pt] # queue
segment = [pt]
while Q: # loop until Q is empty
pt = Q.pop(0) # remove first element of queue
for new_pt in neighbors(pt, shape, structure): # check all neightbors
if data[new_pt] > thres and new_pt not in segment:
Q.append(new_pt)
segment.append(new_pt)
return segment
def find_nconnected(data, pt, thres, diag=False):
"""
Find points connected to pt in data below threshold.
Parameters
----------
data : ndarray
Array of data.
pt : tuple of ints
Starting point of peak.
thres : float
Threshold, above this nodes are considered noise.
diag : bool, optional
True to include diagonal neighbors in connection.
Returns
-------
nodes : list
Indicies of connected nodes.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
if type(pt) == int:
pt = (pt, )
pt = tuple(pt)
shape = data.shape
if data[pt] > thres: # check that the initial point is above threshold.
return []
Q = [pt] # queue
segment = [pt]
while Q: # loop until Q is empty
pt = Q.pop(0) # remove first element of queue
for new_pt in neighbors(pt, shape, structure): # check all neightbors
if data[new_pt] < thres and new_pt not in segment:
Q.append(new_pt)
segment.append(new_pt)
return segment
def find_upward(data, pt, thres, diag=False):
"""
Find points upward-connected to pt in data.
Parameters
----------
data : ndarray
Array of data.
pt : tuple of ints
Starting point of peak.
thres : float
Threshold, below this nodes are considered noise.
diag : bool, optional
True to include diagonal neighbors in connection.
Returns
-------
nodes : list
Indicies of upward-connected nodes.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
if type(pt) == int:
pt = (pt, )
pt = tuple(pt)
shape = data.shape
if data[pt] > thres: # check that the initial point is below threshold.
return []
Q = [pt] # queue
segment = [pt]
while Q: # loop until Q is empty
pt = Q.pop(0) # remove first element of queue
v = data[pt] # value at current node
for new_pt in neighbors(pt, shape, structure): # check all neightbors
if thres > data[new_pt] > v and new_pt not in segment:
Q.append(new_pt)
segment.append(new_pt)
return segment