SynthText数据集简介
SynthText数据集
SynthText in the Wild Dataset
Ankush Gupta, Andrea Vedaldi, and Andrew Zisserman
Visual Geometry Group, University of Oxford, 2016
Data format:
SynthText.zip (size = 42074172 bytes (41GB)) contains 858,750 synthetic
scene-image files (.jpg) split into 200 directories, with
7,266,866 word-instances, and 28,971,487 characters.
Ground-truth annotations are contained in the file “gt.mat” (Matlab format).
The file “gt.mat” contains the following cell-arrays, each of size 1x858750:
-
imnames : names of the image files
-
wordBB : word-level bounding-boxes for each image, represented by
tensors of size 2x4xNWORDS_i, where:
- the first dimension is 2 for x and y respectively,
- the second dimension corresponds to the 4 points
(clockwise, starting from top-left), and
- the third dimension of size NWORDS_i, corresponds to
the number of words in the i_th image. -
charBB : character-level bounding-boxes,
each represented by a tensor of size 2x4xNCHARS_i
(format is same as wordBB’s above) -
txt : text-strings contained in each image (char array).
Words which belong to the same "instance", i.e., those rendered in the same region with the same font, color, distortion etc., are grouped together; the instance boundaries are demarcated by the line-feed character (ASCII: 10) A "word" is any contiguous substring of non-whitespace characters. A "character" is defined as any non-whitespace character.
For any questions or comments, contact Ankush Gupta at:
[email protected]
下载链接
简单介绍一下该数据集，共有80万张图像。两百个图像文件夹，1个gt.mat标注文件。
其中gt.mat文件可以根据以下代码进行数据查看。其中在gt.mat文件中共有1. wordBB(图片中的文本标注)2. charBB(图片中的文本标注) 3. imnames(图片的文本) 4. txt(图片的文本行字符)。四种类型标注。
import scipy
from scipy import io
if __name__ == '__main__':
features_struct = scipy.io.loadmat(r'D:\data\SynText\SynthText\gt.mat')
features = features_struct
# 打印出有多少数据集中共有多少文本字符
print(len(features["wordBB"][0]))
for i in range(len(features["wordBB"][0])):
print(len(features["charBB"][0][i]),features["charBB"][0][i])
print(len(features["wordBB"][0][i]),features["wordBB"][0][i])
print(features["imnames"][0][i])
print(features["txt"][0][i])
print("------------------------------------------------------")
打印出的一列数据
------------------------------------------------------
charBB 15*5+4=79 2*4*79 维度的的字符标注结果 2 是指的坐标 x，y。
4 指的是 共有4个点组成的坐标标注。79是指该图片中共有79个字符。
[[[181.28486148 191.23904421 212.56930465 227.49714618 392.0513942
425.79475759 451.47726272 500.04781649 395.72247711 421.65549013
444.54744753 467.41582437 491.25337715 520.35899958 432.39650645
447.27237987 475.4049677 382.1179307 415.66834197 452.06588854
485.66500521 530.46127148 466.4487681 483.06971896 491.98656061
507.3257657 519.52619265 533.62077208 556.83763058 567.70014145
54.06087905 70.0564337 89.55531722 111.34577116 127.98998409
33.07944351 71.92002348 90.88735442 111.55820797 135.20783299
22.85177068 33.92735782 46.5633709 63.9483445 334.45478996
366.74159856 374.58458318 416.90123067 429.90827245 452.94617151
465.94803099 491.92891522 507.90203733 521.82379256 539.81379198
307.75625592 336.15682495 356.18646505 381.5912669 397.72016039
436.77064033 465.13300429 489.23117218 503.55476533 535.01419565
569.00289901 335.0331587 356.42997237 368.43277742 323.19372016
337.23869626 347.93716773 358.60379034 374.3725247 394.03244882
398.11798414 426.5463602 444.78209025 456.64509425]
[191.28127286 206.30019972 225.10105438 245.39787118 422.81946274
449.58277627 469.29353093 523.75639395 421.6250126 444.54744753
468.40956728 491.25337715 519.03150945 531.27368397 450.3699283
474.19711472 499.32873494 411.85875622 447.9765046 485.05932862
532.13636826 557.78299342 481.98150077 492.68346207 506.30556939
519.60212757 531.68010661 541.03410004 567.70014145 582.03720007
69.84846309 92.26661988 111.34577116 127.98998409 148.40006774
71.06716597 90.88735442 109.26486001 138.20270213 154.47575124
33.63783656 43.65337877 68.06302304 88.11316264 363.57332976
382.80499901 408.65318836 432.92842816 449.92725982 468.9473786
488.93267347 507.90203733 523.86365878 537.77303852 551.76886019
334.01956791 357.36529428 379.36533396 397.72016039 409.80915369
466.89871051 487.22399112 505.28208376 510.58071091 563.00868194
594.95909845 345.31328547 369.40308264 382.31493242 338.30054951
349.03126557 357.50083963 371.26217185 377.50852952 400.19753646
410.37661868 442.51933078 456.64509425 464.49396363]
[204.77989803 220.64720832 236.93684673 261.45180892 423.35510666
450.01362619 469.85339091 524.32142452 422.00662307 444.97084708
468.8763645 491.76167601 519.5902164 531.50604667 450.89279572
474.88718705 499.85377978 440.20431434 478.09096958 517.02823425
566.93275331 594.91385673 483.09465906 493.8059692 507.23218262
520.51818643 532.73155563 542.04654952 569.48786922 584.02303294
62.23683681 84.39705017 106.92491965 124.11568567 145.17626152
61.7306503 85.41840855 104.46173458 132.50363118 149.87639941
17.67934045 29.57931639 54.72446089 80.59555916 363.9616525
383.04260502 409.33066618 433.23585823 450.27575881 469.30488675
489.3179324 508.31361306 524.297361 538.2975689 552.24120686
334.31535104 357.57905011 379.80854094 397.97978941 410.08570643
467.49887373 487.60869951 505.69196636 510.77840796 563.8601076
595.88759267 359.62767303 387.10541681 396.60835088 338.35145348
349.14470964 357.7157564 371.52835133 377.79749479 400.80274091
411.10350155 443.48803694 457.68300726 465.3174714 ]
[194.28783269 204.8349424 223.92219547 242.6754141 392.50877714
426.182479 451.99840093 500.57000724 396.05688377 422.03721225
444.97084708 467.88081522 491.76167601 520.58344701 432.87996505
447.89167201 475.88639051 407.92657929 443.28894518 481.72123344
517.48211666 565.93821345 467.45213208 484.12874003 492.84153013
508.18580656 520.51818643 534.60008213 558.5502169 569.58271409
45.71126117 60.9814573 84.39705017 106.92491965 124.11568567
21.51312702 65.74402385 85.41840855 104.46173458 129.68639926
6.1906517 19.27154164 31.86826505 55.48406412 334.76818112
366.95679406 375.14447365 417.1863514 430.22745901 453.28144286
466.30137219 492.31833244 508.31361306 522.32273706 540.26958872
307.98794183 336.33924719 356.5680906 381.82830223 397.97978941
437.29992369 465.48688826 489.61867974 503.74782401 535.79800256
569.86879207 348.72867785 373.2858834 382.03956434 323.10543247
337.27007863 348.06949256 358.78331622 374.64155801 394.58361913
398.73293753 427.37921041 445.72494419 457.42190203]]
[[288.97105029 282.55288864 288.36855728 285.97493337 13.24681172
13.1839689 8.21544043 13.04568181 64.6938964 66.60126362
66.53110794 66.46102452 66.38797092 81.1768758 115.38859593
109.33174198 120.2061574 252.53386765 256.10788881 256.41691217
254.88830726 253.347558 309.89540703 312.22180576 316.62864702
317.81572079 316.82457426 317.99165328 309.71698972 309.82637127
34.69488047 37.61567872 54.89471148 64.25653825 71.4073869
67.15538757 91.91051319 99.35514807 98.56014348 116.75083658
254.97315834 256.58855769 254.96873897 267.32733873 134.86364826
148.78895027 132.67157984 148.5427343 147.47409608 148.36580238
148.30198083 148.17444982 148.09604339 146.02243073 147.93939983
199.64668986 212.67581014 196.31262481 213.40318887 213.30055323
199.87916866 212.87157398 212.71822614 225.80556005 198.2731329
198.06770368 253.48007711 262.77438083 259.87187523 301.63463255
308.64871588 304.20386424 308.85485828 306.75324085 303.6109472
302.54642675 302.89238716 309.68635479 309.80081569]
[287.19711992 279.61868489 286.35913778 283.10461864 13.18951001
13.13966667 8.18423461 13.00152753 64.61567217 66.53110794
66.45797905 66.38797092 66.30284086 81.13976259 115.31375695
109.22326643 120.10384785 252.35019484 255.8259365 256.11467832
254.50278521 253.14426046 310.04527554 312.30645769 316.7305178
317.89785847 316.91104175 318.04125378 309.82637127 309.97074027
41.54182107 47.32221154 64.25653825 71.4073869 80.17616556
82.35689029 99.35514807 106.56827783 109.22243972 124.31345131
254.97114803 256.5858348 254.96473187 267.30552222 134.72992899
148.7101007 132.51666682 148.46406244 147.3762828 148.28725809
148.18915732 148.09604339 148.01769341 145.94486306 147.88071661
199.48795325 212.54133515 196.17411524 213.30055323 213.22362535
199.69638655 212.73099876 212.61608673 225.75876259 198.10393299
197.91082328 253.41227785 262.60940711 259.72030215 301.81847579
308.76249719 304.31222712 308.97699343 306.78613864 303.68338879
302.69560876 303.08677094 309.80081569 309.87654598]
[317.10536435 311.44071295 312.63529648 318.7904856 47.75629342
37.7972408 37.74927224 37.61757951 89.47734975 91.38709268
91.30044754 91.21750012 91.11663605 91.07344733 145.36337712
145.24722529 145.12636724 280.30964758 282.39090725 281.31532312
278.50004401 277.06052901 329.68692661 330.80352767 330.82471226
330.84567561 330.86494651 330.8796442 329.86022974 330.94587691
60.09070207 69.30143095 78.66497338 85.81018821 94.56386089
107.25079324 116.16992694 123.34029609 133.89889021 147.88938016
287.57468436 286.64010828 286.58936313 287.44440244 169.96893263
167.84838286 179.81667877 167.5803956 167.48941772 167.38781913
167.28096729 167.17954728 167.09420831 168.02476669 166.94501307
233.04701874 232.88887147 232.73775478 232.61422614 232.53192978
232.14163349 232.00492634 231.88199579 234.89444122 231.48656717
231.26884344 272.17233118 284.08629122 275.96890498 325.91767514
327.10717246 327.13169518 327.17121455 323.76683086 327.25497184
327.28444346 326.26337629 331.85808581 326.33515397]
[317.76304835 312.68024838 313.55295862 319.67662241 47.83774432
37.85485971 37.79244201 37.67500565 89.57023995 91.47021524
91.38709268 91.30405589 91.21750012 91.11303611 145.45057346
145.37790493 145.24238836 281.13200847 283.35226151 282.2906266
279.81266027 277.81416154 329.655548 330.78825842 330.80200592
330.82621693 330.84567561 330.86789477 329.83828902 330.92309221
53.22196871 59.56891285 69.30143095 78.66497338 85.81018821
92.10771961 108.76194575 116.16992694 123.34029609 140.42118448
287.59847423 286.66091029 286.63548898 287.49640134 170.12612462
167.93426675 180.00851812 167.66608567 167.59645777 167.47337004
167.40385522 167.26494783 167.17954728 168.11042051 167.00893084
233.22599319 233.03326024 232.895744 232.72402439 232.61422614
232.34692678 232.15531102 231.99126244 234.94271852 231.67733413
231.44571997 272.37780781 284.46055286 276.28717833 325.86754448
327.07319781 327.10409614 327.13474958 323.75376213 327.23717824
327.2490499 326.21040862 331.84434182 326.30919243]]]
# 2*4*19 和上述同理
[[[181.6932 389.59775 500.04785 395.7225 432.3676
382.11975 466.44873 556.8377 54.828644 34.546387
6.1231613 334.4438 307.73584 436.82367 535.0142
335.33475 323.19373 394.00983 426.54636 ]
[262.91235 469.1694 523.7898 531.4252 499.67932
595.04297 542.24194 584.05646 154.29362 162.3328
88.08734 552.04333 409.83698 512.0845 595.66046
398.89645 378.05872 410.51965 465.59695 ]
[261.38748 472.29318 524.3214 531.5062 499.85382
594.87524 542.04205 584.023 145.17624 149.29953
88.15483 552.30896 410.0857 510.7658 595.8875
394.54288 377.75128 411.10352 465.24753 ]
[180.16832 392.72153 500.57947 395.80362 432.5421
381.95203 466.2488 556.80426 45.71126 21.513126
6.190651 334.7094 307.98456 435.505 535.2413
330.98117 322.8863 394.5937 426.19693 ]]
[[278.64554 13.454558 13.045681 64.69389 109.40389
252.18835 309.89542 309.71695 31.286362 62.926376
255.0078 132.89682 196.64201 198.5215 198.27312
251.71115 301.63464 302.64227 302.8924 ]
[281.8575 6.717518 12.542309 64.28407 109.078064
253.34705 310.62674 309.75986 72.62825 119.99968
254.83804 131.67578 195.94792 201.46156 197.86086
262.54715 302.3023 302.25714 303.36755 ]
[320.41666 43.612892 37.61758 91.13962 145.12637
284.1772 331.34708 330.94583 94.56386 149.18102
287.4287 179.01437 232.53194 235.21687 231.26884
288.08444 327.56586 327.2845 332.08185 ]
[317.20468 50.349934 38.120953 91.54944 145.45221
283.0185 330.61575 330.90292 53.22197 92.10771
287.59848 180.23541 233.22603 232.27681 231.6811
277.24844 326.8982 327.66962 331.6067 ]]]
图片名称
['8/ballet_26_58.jpg']
# 图片中的字符，可以查一下 共计19个单词 79个字符。
['they ' 'get a \n>scan.\n the '
'>come ' 'Lines: 15 '
'there\nX11R5 ' 'like '
' References: \nDate: Tue, 20' 'the '
'Date: 19 Apr ']
数据正在上传百度云盘中，如需链接后续将会放出。
或直接联系邮箱[email protected]