forked from idhmc-tamu/eMOP
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathocr_toolkit.py
610 lines (495 loc) · 20.5 KB
/
ocr_toolkit.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
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
#
# Copyright (C) 2009-2010 Rene Baston, Christoph Dalitz
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
#
from gamera.core import *
init_gamera()
from gamera import knn
from gamera.plugins import pagesegmentation
from gamera.classify import ShapedGroupingFunction
from gamera.plugins.image_utilities import union_images
from gamera.plugins.listutilities import median
from gamera.toolkits.ocr.classes import Textline
import unicodedata
import sys
import time
def return_char(unicode_str, extra_chars_dict={}):
"""Converts a unicode character name to a unicode symbol.
Signature:
``return_char (classname, extra_chars_dict={})``
with
*classname*:
A class name derived from a unicode character name.
Example: ``latin.small.letter.a`` returns the character ``a``.
*extra_chars_dict*
A dictionary of additional translations of classnames to character codes.
This is necessary when you use class names that are not unicode names.
The character 'code' does not need to be an actual code, but can be
any string. This can be useful, e.g. for ligatures:
.. code:: Python
return_char(glyph.get_main_id(), {'latin.small.ligature.st':'st'})
When *classname* is not listed in *extra_chars_dict*, it must correspond
to a `standard unicode character name`_,
as in the examples of the following table:
.. _`standard unicode character names`: http://www.unicode.org/charts/
+-----------+----------------------------+----------------------------+
| Character | Unicode Name | Class Name |
+===========+============================+============================+
| ``!`` | ``EXCLAMATION MARK`` | ``exclamation.mark`` |
+-----------+----------------------------+----------------------------+
| ``2`` | ``DIGIT TWO`` | ``digit.two`` |
+-----------+----------------------------+----------------------------+
| ``A`` | ``LATIN CAPITAL LETTER A`` | ``latin.capital.letter.a`` |
+-----------+----------------------------+----------------------------+
| ``a`` | ``LATIN SMALL LETTER A`` | ``latin.small.letter.a`` |
+-----------+----------------------------+----------------------------+
"""
if len(extra_chars_dict) > 0:
try:
return extra_chars_dict[unicode_str]
except:
pass
name = unicode_str.upper()
# some xml-files might be corrupted due to wrong grouping
if name.startswith('_GROUP.'):
name = name[len('_GROUP.'):]
if name.startswith('_PART.'):
name = name[len('_PART.'):]
name = name.replace(".", " ")
try:
return unicodedata.lookup(name)
except KeyError:
strings = unicode_str.split(".")
if(strings[0] == "collated"):
return strings[1]
if(strings[0] == "cursive"):
return return_char(unicode_str[8:])
else:
print "ERROR: Name not found:", name
return ""
def chars_make_words(lines_glyphs,threshold=None):
"""Groups the given glyphs to words based upon the horizontal distance
between adjacent glyphs.
Signature:
``chars_make_words (glyphs, threshold=None)``
with
*glyphs*:
A list of ``Cc`` data types, each of which representing a character.
All glyphs must stem from the same single line of text.
*threshold*:
Horizontal white space greater than *threshold* will be considered
a word separating gap. When ``None``, the threshold value is
calculated automatically as 2.5 times teh median white space
between adjacent glyphs.
The result is a nested list of glyphs with each sublist representing
a word. This is the same data structure as used in `Textline.words`_
.. _`Textline.words`: gamera.toolkits.ocr.classes.Textline.html
"""
glyphs = lines_glyphs[:]
wordlist = []
if(threshold == None):
spacelist = []
total_space = 0
for i in range(len(glyphs) - 1):
spacelist.append(glyphs[i + 1].ul_x - glyphs[i].lr_x)
if(len(spacelist) > 0):
threshold = median(spacelist)
threshold = threshold * 2.5
else:
threshold = 0
word = []
for i in range(len(glyphs)):
if i > 0:
if((glyphs[i].ul_x - glyphs[i - 1].lr_x) > threshold):
wordlist.append(word)
word = []
word.append(glyphs[i])
if(len(word) > 0):
wordlist.append(word)
return wordlist
def textline_to_string(line, heuristic_rules="roman", extra_chars_dict={}):
"""Returns a unicode string of the text in the given ``Textline``.
Signature:
``textline_to_string (textline, heuristic_rules="roman", extra_chars_dict={})``
with
*textline*:
A ``Textline`` object containing the glyphs. The glyphs must already
be classified.
*heuristic_rules*:
Depending on the alphabeth, some characters can very similar and
need further heuristic rules for disambiguation, like apostroph and
comma, which have the same shape and only differ in their position
relative to the baseline.
When set to \"roman\", several rules specific for latin alphabeths
are applied.
*extra_chars_dict*
A dictionary of additional translations of classnames to character codes.
This is necessary when you use class names that are not unicode names.
Will be passed to `return_char`_.
As this function uses `return_char`_, the class names of the glyphs in
*textline* must corerspond to unicode character names, as described in
the documentation of `return_char`_.
.. _`return_char`: #return-char
"""
wordlist = line.words
s = ""
char = ""
for i in range(len(wordlist)):
if(i):
s = s + " "
for glyph in wordlist[i]:
char = return_char(glyph.get_main_id(), extra_chars_dict)
if (heuristic_rules == "roman"):
# disambiguation of similar roman characters
if (char == "x" or char == "X"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.x")
else:
glyph.classify_heuristic("latin.small.letter.x")
char = return_char(glyph.get_main_id())
if (char == "p" or char == "P"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.p")
else:
glyph.classify_heuristic("latin.small.letter.p")
char = return_char(glyph.get_main_id())
if (char == "o" or char == "O"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.o")
else:
glyph.classify_heuristic("latin.small.letter.o")
char = return_char(glyph.get_main_id())
if (char == "w" or char == "W"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.w")
else:
glyph.classify_heuristic("latin.small.letter.w")
char = return_char(glyph.get_main_id())
if (char == "v" or char == "V"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.v")
else:
glyph.classify_heuristic("latin.small.letter.v")
char = return_char(glyph.get_main_id())
if (char == "z" or char == "Z"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.z")
else:
glyph.classify_heuristic("latin.small.letter.z")
char = return_char(glyph.get_main_id())
if (char == "s" or char == "S"):
# not for long s
if (glyph.get_main_id().upper() != "LATIN.SMALL.LETTER.LONG.S"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.s")
else:
glyph.classify_heuristic("latin.small.letter.s")
char = return_char(glyph.get_main_id())
#if(char == "T" and (float(glyph.nrows)/float(glyph.ncols)) > 1.5):
# glyph.classify_heuristic("LATIN SMALL LETTER F")
# char = return_char(glyph.get_main_id())
if (char == "'" or char == ","):
if (glyph.ul_y < line.bbox.center_y):
glyph.classify_heuristic("APOSTROPHE")
char = "'"
else:
glyph.classify_heuristic("COMMA")
char = ","
s = s + char
return s
def textline_to_xml(line, heuristic_rules="roman", extra_chars_dict={}):
"""Returns xml encoding of words and coordinates for the text in the given ``Textline``.
Signature:
``textline_to_xml (textline, heuristic_rules="roman", extra_chars_dict={})``
with
*textline*:
A ``Textline`` object containing the glyphs. The glyphs must already
be classified.
*heuristic_rules*:
Depending on the alphabeth, some characters can very similar and
need further heuristic rules for disambiguation, like apostroph and
comma, which have the same shape and only differ in their position
relative to the baseline.
When set to \"roman\", several rules specific for latin alphabeths
are applied.
*extra_chars_dict*
A dictionary of additional translations of classnames to character codes.
This is necessary when you use class names that are not unicode names.
Will be passed to `return_char`_.
As this function uses `return_char`_, the class names of the glyphs in
*textline* must corerspond to unicode character names, as described in
the documentation of `return_char`_.
.. _`return_char`: #return-char
"""
# This function was added by Dave Woods - woodsdm2@muohio.edu in 9/2010
# It is based on the textline_to_string function, but modified to produce output with XML tags
# These tags add tagging of words, along with the coordinates for the upper right and lower
# left corners of the word bounding box.
# Added to support the requirements of the 18th Connect project.
#
# Modified by Matt Christy - mchristy@tamu.edu on 7/24/2013
# Correcting to more accurately reflect the Gale OCR XML structure:
# chainge <line> to <p>
wordlist = line.words
s = "<p>\n"
char = ""
for i in range(len(wordlist)):
word = ""
#set left/right x and upper/lower y from first glyph
word_leftx = wordlist[i][0].ul_x
word_uppery = wordlist[i][0].ul_y
word_rightx = wordlist[i][0].lr_x
word_lowery = wordlist[i][0].lr_y
for glyph in wordlist[i]:
#update right x and (conditionally) upper/lower y from current glyph
word_rightx = glyph.lr_x
if (glyph.ul_y < word_uppery):
word_uppery = glyph.ul_y
if (glyph.lr_y > word_lowery):
word_lowery = glyph.lr_y
char = return_char(glyph.get_main_id(), extra_chars_dict)
if (heuristic_rules == "roman"):
# disambiguation of similar roman characters
if (char == "x" or char == "X"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.x")
else:
glyph.classify_heuristic("latin.small.letter.x")
char = return_char(glyph.get_main_id())
if (char == "p" or char == "P"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.p")
else:
glyph.classify_heuristic("latin.small.letter.p")
char = return_char(glyph.get_main_id())
if (char == "o" or char == "O"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.o")
else:
glyph.classify_heuristic("latin.small.letter.o")
char = return_char(glyph.get_main_id())
if (char == "w" or char == "W"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.w")
else:
glyph.classify_heuristic("latin.small.letter.w")
char = return_char(glyph.get_main_id())
if (char == "v" or char == "V"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.v")
else:
glyph.classify_heuristic("latin.small.letter.v")
char = return_char(glyph.get_main_id())
if (char == "z" or char == "Z"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.z")
else:
glyph.classify_heuristic("latin.small.letter.z")
char = return_char(glyph.get_main_id())
if (char == "s" or char == "S"):
# not for long s
if (glyph.get_main_id().upper() != "LATIN.SMALL.LETTER.LONG.S"):
if (glyph.ul_y <= line.bbox.center_y-(line.bbox.nrows/4)):
glyph.classify_heuristic("latin.capital.letter.s")
else:
glyph.classify_heuristic("latin.small.letter.s")
char = return_char(glyph.get_main_id())
#if(char == "T" and (float(glyph.nrows)/float(glyph.ncols)) > 1.5):
# glyph.classify_heuristic("LATIN SMALL LETTER F")
# char = return_char(glyph.get_main_id())
if (char == "'" or char == ","):
if (glyph.ul_y < line.bbox.center_y):
glyph.classify_heuristic("APOSTROPHE")
char = "'"
else:
glyph.classify_heuristic("COMMA")
char = ","
# Handle XML predefined entities - &, ', ", <, and >
if (char == '&'):
char = '&'
if (char == "'"):
char = '''
if (char == '"'):
char = '"'
if (char == '<'):
char = '<'
if (char == '>'):
char = '>'
word = word + char
# end of glyph processing for word
pos = "pos=\"" +str(word_leftx)+','+str(word_uppery)+','+str(word_rightx)+','+str(word_lowery)+'"'
s = s + "<wd "+pos+">"+word+"</wd>\n"
s = s + "</p>\n"
return s
def check_upper_neighbors(item,glyph,line):
"""Check for small signs grouped beside each other like quotation marks.
Signature:
``check_upper_neighbors(item,glyph,line)``
with
*item*:
Some connected-component.
*glyph*:
Some connected-component.
*line*:
The ``Textline`` Object which includes ``item`` and ``glyph``
Returns an array with two elements. The first element keeps a list of
characters (images that has been united to a single image) and the
second image is a list of characters which has to be removed as
these have been united to a single character.
"""
remove = []
add = []
result = []
minheight = min([item.nrows,glyph.nrows])
# glyphs must be small, of similar size and on the same height
if(not(glyph.lr_y >= line.center_y and glyph.lr_y-(glyph.nrows/3) <= line.lr_y)):
if (glyph.contains_y(item.center_y) and item.contains_y(glyph.center_y)):
minwidth = min([item.ncols,glyph.ncols])
distance = item.lr_x - glyph.lr_x
if(distance > 0 and distance <= minwidth*3):
remove.append(item)
remove.append(glyph)
new = union_images([item,glyph])
add.append(new)
result.append(add) #result[0] == ADD
result.append(remove) #result[1] == REMOVE
return result
def check_glyph_accent(item,glyph):
"""Check two glyphs for beeing grouped to one single character. This function is for unit connected-components like i, j or colon.
Signature:
``check_glyph_accent(item,glyph)``
with
*item*:
Some connected-component.
*glyph*:
Some connected-component.
There is returned an array with two elements. The first element keeps a list of characters (images that has been united to a single image) and the second image is a list of characters which has to be removed as these have been united to a single character.
"""
remove = []
add = []
result = []
if(glyph.contains_x(item.ul_x) or glyph.contains_x(item.lr_x) or glyph.contains_x(item.center_x)): ##nebeinander?
if(not(item.contains_y(glyph.ul_y) or item.contains_y(glyph.lr_y) or item.contains_y(glyph.center_y))): ##nicht y-dimensions ueberschneident
remove.append(item)
remove.append(glyph)
new = union_images([item,glyph])
add.append(new)
result.append(add) #result[0] == ADD
result.append(remove) #result[1] == REMOVE
return result
def get_line_glyphs(image,textlines):
"""Splits image regions representing text lines into characters.
Signature:
``get_line_glyphs (image, segments)``
with
*image*:
The document image that is to be further segmentated. It must contin the
same underlying image data as the second argument *segments*
*segments*:
A list ``Cc`` data types, each of which represents a text line region.
The image views must correspond to *image*, i.e. each pixels has a value
that is the unique label of the text line it belongs to. This is the
interface used by the plugins in the \"PageSegmentation\" section of the
Gamera core.
The result is returned as a list of Textline_ objects.
.. _Textline: gamera.toolkits.ocr.classes.Textline.html
"""
i=0
show = []
lines = []
ret,sub_ccs = image.sub_cc_analysis(textlines)
for ccs in sub_ccs:
line_bbox = Rect(textlines[i])
i = i + 1
glyphs = ccs[:]
newlist = []
remove = []
add = []
result = []
glyphs.sort(lambda x,y: cmp(x.ul_x, y.ul_x))
for position, item in enumerate(glyphs):
if(True):
#if(not(glyph.lr_y >= line_bbox.center_y and glyph.lr_y-(glyph.nrows/3) <= line_bbox.lr_y)): ## is this part of glyph higher then line.center_y ?
left = position - 2
if(left < 0):
left = 0
right = position + 2
if(right > len(glyphs)):
right = len(glyphs)
checklist = glyphs[left:right]
for glyph in checklist:
if (item == glyph):
continue
result = check_upper_neighbors(glyph,item,line_bbox)
if(len(result[0]) > 0): #something has been joind...
joind_upper_connection = result[0][0] #joind glyph
add.append(joind_upper_connection)
remove.append(result[1][0]) #first part of joind one
remove.append(result[1][1]) #second part of joind one
for glyph2 in checklist: #maybe the upper joind glyphs fits to a glyph below...
if(glyphs == joind_upper_connection):
continue
if(joind_upper_connection.contains_x(glyph2.center_x)): #fits for example on ae, oe, ue in german alph
new = union_images([glyph2,joind_upper_connection])
add.append(new)
remove.append(glyph2)
add.remove(joind_upper_connection)
break
for elem in remove:
if (elem in checklist):
checklist.remove(elem)
for glyph in checklist:
if(item == glyph):
continue
result = check_glyph_accent(item,glyph)
if(len(result[0]) > 0): #something has been joind...
add.append(result[0][0]) #joind glyph
remove.append(result[1][0]) #first part of joind one
remove.append(result[1][1]) #second part of joind one
for elem in remove:
if(elem in glyphs):
glyphs.remove(elem)
for elem in add:
glyphs.append(elem)
new_line = Textline(line_bbox)
final = []
if(len(glyphs) > 0):
for glyph in glyphs:
final.append(glyph)
new_line.add_glyphs(final,False)
new_line.sort_glyphs() #reading order -- from left to right
lines.append(new_line)
for glyph in glyphs:
show.append(glyph)
return lines
def show_bboxes(image,glyphs):
"""Returns an RGB image with bounding boxes of the given glyphs as
hollow rects. Useful for visualization and debugging of a segmentation.
Signature:
``show_bboxes (image, glyphs)``
with:
*image*:
An image of the textdokument which has to be segmentated.
*glyphs*:
List of rects which will be drawn on ``image`` as hollow rects.
As all image types are derived from ``Rect``, any image list can
be passed.
"""
rgb = image.to_rgb()
if(len(glyphs) > 0):
for glyph in glyphs:
rgb.draw_hollow_rect(glyph, RGBPixel(255,0,0), 1.0)
return rgb