-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathvqa_v1_download.html
595 lines (590 loc) · 38.5 KB
/
vqa_v1_download.html
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
<!DOCTYPE html>
<html lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>VQA: Visual Question Answering</title>
<link rel="stylesheet" href="./static/css/foundation.css">
<link rel="stylesheet" href="./static/css/main.css">
</head>
<body class="off-canvas hide-extras" style="min-width:1200px; min-height:750px;">
<header>
<div class="row">
<a href="http://visualqa.org/"><img style="height: 100px; position:absolute; top:4px; left:0px;" src="./static/img/main.png" alt="logo" /></a>
<h1><img style="height: 90px;" src="./static/img/logo.png" alt="logo" /><br></h1>
<br>
</div>
</header>
<div class="contain-to-grid">
<nav class="top-bar" data-topbar>
<section class="top-bar-section">
<!-- Right Nav Section -->
<ul class="right">
<li><a href="index.html">Home</a></li>
<li><a href="people.html">People</a></li>
<li><a href="code.html">Code</a></li>
<li><a href="http://vqa.cloudcv.org/" onClick="ga('send', 'event', { eventCategory: 'Outgoing Link', eventAction: 'Demo', eventLabel: 'Demo'});">Demo</a></li>
<li class="active has-dropdown"><a href="download.html">Download</a>
<ul class="dropdown">
<li><a href="download.html">VQA v2</a></li>
<li><a href="vqa_v1_download.html">VQA v1</a></li>
</ul>
</li>
<li><a href="evaluation.html">Evaluation</a></li>
<li class="has-dropdown"><a href="challenge.html">Challenge</a>
<ul class="dropdown">
<li><a href="challenge.html">2021</a></li>
<li><a href="challenge_2020.html">2020</a></li>
<li><a href="challenge_2019.html">2019</a></li><li><a href="challenge_2018.html">2018</a></li>
<li><a href="challenge_2017.html">2017</a></li>
<li><a href="challenge_2016.html">2016</a></li>
</ul>
</li>
<li class="has-dropdown"><a href="http://visualqa.org/vqa_v2_teaser.html">Browse</a>
<ul class="dropdown">
<li><a href="http://visualqa.org/vqa_v2_teaser.html">VQA v2</a></li>
<li><a href="https://vqabrowser.cloudcv.org/">VQA v1</a></li>
</ul>
</li>
<li><a href="http://visualqa.org/visualize/">Visualize</a></li>
<!-- <li class="has-dropdown"><a href="http://visualqa.org/visualize/">Visualize</a>
<ul class="dropdown">
<li><a href="http://visualqa.org/visualize/">VQA v2</a></li>
<li><a href="http://visualqa.org/visualize/">VQA v1</a></li>
</ul>
</li> -->
<li class="has-dropdown"><a href="workshop.html">Workshop</a>
<ul class="dropdown">
<li><a href="workshop.html">2021</a></li>
<li><a href="workshop_2020.html">2020</a></li>
<li><a href="workshop_2019.html">2019</a></li>
<li><a href="workshop_2018.html">2018</a></li>
<li><a href="workshop_2017.html">2017</a></li>
<li><a href="workshop_2016.html">2016</a></li>
</ul>
</li>
<li><a href="sponsors.html">Sponsors</a></li>
<li><a href="terms.html">Terms</a></li>
<li><a href="external.html">External</a></li>
</ul>
</section>
</nav>
</div>
<section role="main" style="padding: 1em;">
<div class="row">
<h1 style="font-size:30px; color:grey; font-weight: 200; text-align:left;">Download</h1>
<div class="large-3 columns">
<div class="row">
<h1 style="font-size:19px; color:grey; font-weight:200;" align=left>VQA Annotations</h1></div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight:200; text-align:left;">Real Images</h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns"> </div>
</div>
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/mscoco/vqa/Annotations_Train_mscoco.zip" onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Annotations', eventLabel: 'Train_Annotations_Real'});">Training annotations 2015 v1.0</a>
<br>
<span style="font-size: 13px">2,483,490 answers</span>
</li>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/mscoco/vqa/Annotations_Val_mscoco.zip" onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Annotations', eventLabel: 'Val_Annotations_Real'});">Validation annotations 2015 v1.0</a>
<br>
<span style="font-size:13px">1,215,120 answers</span>
</li>
<li style="color: white;"><span style="font-size: 15px"> </span>
<br><span style="font-size:13px"> </span></li>
</ul>
</div>
</div>
</div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight:200; text-align:left;">Abstract Scenes</h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns"> </div>
</div>
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/vqa/Annotations_Train_abstract_v002.zip" onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Annotations', eventLabel: 'Train_Annotations_Abstract'});">Training annotations 2015 v1.0</a>
<br>
<span style="font-size: 13px">600,000 answers</span>
</li>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/vqa/Annotations_Val_abstract_v002.zip" onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Annotations', eventLabel: 'Val_Annotations_Abstract'});">Validation annotations 2015 v1.0</a>
<br>
<span style="font-size: 13px">300,000 answers</span>
</li>
</ul>
</div>
</div>
</div>
</div>
<div class="large-3 columns">
<div class="row">
<h1 style="font-size:19px; color:grey; font-weight:200;" align=left>VQA Input Questions</h1></div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight:200; text-align:left;"> </h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns"> </div>
</div>
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li><a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/mscoco/vqa/Questions_Train_mscoco.zip" onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Questions', eventLabel: 'Train_Questions_Real'});">Training questions 2015 v1.0</a>
<br>
<span style="font-size: 13px">248,349 questions</span>
</li>
<li><a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/mscoco/vqa/Questions_Val_mscoco.zip" onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Questions', eventLabel: 'Val_Questions_Real'});">Validation questions 2015 v1.0</a>
<br>
<span style="font-size: 13px">121,512 questions</span>
</li>
<li><a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/mscoco/vqa/Questions_Test_mscoco.zip" onClick="ga('send', 'event', { eventCategory: 'Test Set Download', eventAction: 'Testing Questions', eventLabel: 'Test_Questions_Real'});">Testing questions 2015 v1.0</a>
<br>
<span style="font-size: 13px">244,302 questions</span>
</li>
</ul>
</div>
</div>
</div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight:200; text-align:left;"> </h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns"> </div>
</div>
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li><a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/vqa/Questions_Train_abstract_v002.zip" onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Questions', eventLabel: 'Train_Questions_Abstract'});">Training questions 2015 v1.0</a>
<br>
<span style="font-size: 13px">60,000 questions</span>
</li>
<li><a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/vqa/Questions_Val_abstract_v002.zip" onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Questions', eventLabel: 'Val_Questions_Abstract'});">Validation questions 2015 v1.0</a>
<br>
<span style="font-size: 13px">30,000 questions</span>
</li>
<li><a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/vqa/Questions_Test_abstract_v002.zip" onClick="ga('send', 'event', { eventCategory: 'Testing Set Download', eventAction: 'Testing Questions', eventLabel: 'Test_Questions_Abstract'});">Testing questions 2015 v1.0</a>
<br>
<span style="font-size: 13px">60,000 questions</span>
</li>
</ul>
</div>
</div>
</div>
</div>
<div class="large-2 columns" style="margin-left:0px;">
<div class="row">
<h1 style="font-size:19px; color:grey; font-weight:200;" align=left>VQA Input Images</h1></div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight: 200; text-align:left;"> </h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns" style="margin-left:30px;"><a href="http://mscoco.org/dataset/#download">MSCOCO</a></div>
</div>
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li>
<a style="font-size: 15px"
href="http://images.cocodataset.org/zips/train2014.zip"
onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Images', eventLabel: 'Train_Images_Real'});">Training images</a>
<br><span style="font-size: 13px">82,783 images</span>
</li>
<li>
<a style="font-size: 15px"
href="http://images.cocodataset.org/zips/val2014.zip"
onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Images', eventLabel: 'Val_Images_Real'});">Validation images</a>
<br><span style="font-size: 13px">40,504 images</span>
</li>
<li>
<a style="font-size: 15px"
href="http://images.cocodataset.org/zips/test2015.zip"
onClick="ga('send', 'event', { eventCategory: 'Test Set Download', eventAction: 'Testing Images', eventLabel: 'Test_Images_Real'});">Testing images</a>
<br><span style="font-size: 13px">81,434 images</span>
</li>
</ul>
</div>
</div>
</div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight: 200; text-align:left;"> </h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns"> </div>
</div>
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/scene_img/scene_img_abstract_v002_train2015.zip" onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Images', eventLabel: 'Train_Images_Abstract'});">Training images</a>
<br><span style="font-size: 13px">20,000 images</span>
</li>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/scene_img/scene_img_abstract_v002_val2015.zip" onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Images', eventLabel: 'Val_Images_Abstract'});">Validation images</a>
<br><span style="font-size: 13px">10,000 images</span>
</li>
<li>
<a style="font-size: 15px" href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/scene_img/scene_img_abstract_v002_test2015.zip" onClick="ga('send', 'event', { eventCategory: 'Testing Set Download', eventAction: 'Testing Images', eventLabel: 'Test_Images_Abstract'});">Testing images</a>
<br><span style="font-size: 13px">20,000 images</span>
</li>
</ul>
</div>
</div>
</div>
</div>
<div class="large-3 columns" style="margin-left:0px;">
<div class="row">
<h1 style="font-size:19px; color:grey; font-weight: 200;" align=left>Tools (Real & Abstract)</h1></div>
<div class="row">
<h1 style="font-size:18px; color:grey; font-weight: 200; text-align:left;"> </h1>
<div class="row" style="font-size:15px;">
<div class="large-12 columns"> </div>
</div>
<div class="row">
<div class="large-12 columns" style="margin-top:0px;">
<div class="large-12 columns">
<ul>
<li><a style="font-size: 15px" href="https://github.com/VT-vision-lab/VQA">VQA API</a></li>
<li><a style="font-size: 15px" href="https://github.com/VT-vision-lab/VQA/tree/master/PythonEvaluationTools">Evaluation Code</a></li>
<li><a style="font-size: 15px" href="https://github.com/VT-vision-lab/abstract_scenes_v002">Abstract Scenes Code</a></li>
</ul>
</div>
</div>
</div>
</div>
</div>
<div class="row">
<div class="large-11 columns">
<h1 style="font-size:15px; color:black; font-weight: 200; text-align:left;">The captions for training and validation sets of the abstract scenes can be downloaded from
<a href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/captions_abstract_v002_train-val2015.zip" onClick="ga('send', 'event', { eventCategory: 'Trainval Download', eventAction: 'Trainval Captions', eventLabel: 'Trainval_Captions_Abstract'});">here</a>.</h1>
</div>
</div>
<hr>
</div>
<!-- End of the download table thing -->
<div class="row">
<h1 style="font-size:30px; color:grey; font-weight: 200; text-align:left;">Overview</h1>
<div class="large-12 columns">
<p style="font-size:15px; font-weight: 200; display: inline;" align=left>
For every image, we collected 3 free-form natural-language questions with 10 concise open-ended answers each. We provide two formats of the VQA task: open-ended and multiple-choice. For additional details, please see the <a href="http://arxiv.org/pdf/1505.00468v6.pdf" onClick="ga('send', 'event', { eventCategory: 'ICCV2015 Paper Download', eventAction: 'Paper', eventLabel: 'ICCV2015_Paper'});">VQA paper</a>.
<br>
<br> The annotations we release are the result of the following post-processing steps on the raw crowdsourced data:</p>
<div class="large-12 columns">
<ul style="font-size:15px; font-weight: 200; display: inline;" align=left>
<li>Spelling correction (using Bing Speller) of question and answer strings</li>
<li>Question normalization (first char uppercase, last char ‘?’)</li>
<li>Answer normalization (all chars lowercase, no period except as decimal point, number words —> digits, strip articles (a, an the))</li>
<li>Adding apostrophe if a contraction is missing it (e.g., convert "dont" to "don't")</li>
</ul>
<br>
</div>
</div>
<div class="large-12 columns">
<p style="font-size:15px; font-weight: 200; text-align:left">Please follow the instructions in the <a href="https://github.com/VT-vision-lab/VQA/blob/master/README.md" target="_blank">README</a> to download and setup the VQA data (annotations and images).
<br> By downloading this dataset, you agree to our <a href="terms.html">Terms of Use</a>.
</p>
</div>
<hr>
</div>
<div class="row">
<h1 style="font-size:30px; color:grey; font-weight: 200; text-align:left;">VQA API</h1>
<div class="large-12 columns" style="text-align:left;">
<p style="font-size:15px; font-weight: 200; text-align:left; border-style: solid; padding-left:5px; border-width: 1px;"><code>getQuesIds - Get question ids that satisfy given filter conditions.<br>
getImgIds - Get image ids that satisfy given filter conditions.<br>
loadQA - Load questions and answers with the specified question ids.<br>
showQA - Display the specified questions and answers.<br>
loadRes - Load result file and create result object.</code></p>
<p style="font-size:15px; font-weight: 200;">Here is a <a href="https://github.com/VT-vision-lab/VQA/blob/master/PythonHelperTools/vqaDemo.py" target="_blank">link</a> to the python API demo script.</p>
</div>
<hr>
</div>
<div class="row">
<h1 style="font-size:30px; color:grey; font-weight: 200">Input Questions Format</h1>
<div class="large-12 columns" style="text-align:center;">
<p style="font-size:15px; font-weight: 200; text-align:left;">VQA currently has two different question formats: OpenEnded and MultipleChoice. The questions are stored using the JSON file format.
<br>
<br> The OpenEnded format has the following data structure:
<br>
</p>
<p style="font-size:15px; font-weight: 200; border-style: solid;
border-width: 1px; text-align:left; padding:5px;"><code>{<br>
"info" : info,<br>
"task_type" : str,<br>
"data_type": str,<br>
"data_subtype": str,<br>
"questions" : [question],<br>
"license" : license<br>
}<br>
<br>
info {<br>
"year" : int,<br>
"version" : str,<br>
"description" : str,<br>
"contributor" : str,<br>
"url" : str,<br>
"date_created" : datetime<br>
}<br>
<br>
license{<br>
"name" : str,<br>
"url" : str<br>
}<br>
<br>
question{<br>
"question_id" : int,<br>
"image_id" : int,<br>
"question" : str<br>
}<br></code></p>
<p style="font-size:15px; font-weight: 200; text-align:left;">The MultipleChoice format has the same data structure as the OpenEnded format above, except it has the following two extra fields:</p>
<p style="font-size:15px; font-weight: 200; border-style: solid;
border-width: 1px; text-align:justify; padding:20px;"><code>"num_choices": int<br>
question{<br>
"multiple_choices" : [str]<br>
}</code></p>
<p style="font-size:15px; font-weight: 200; text-align:left;"><code>task_type</code>: type of annotations in the JSON file (OpenEnded/MultipleChoice).
<br>
<code>data_type</code>: source of the images (mscoco or abstract_v002).
<br>
<code>data_subtype</code>: type of data subtype (train2014/val2014/test2015/test-dev2015 for mscoco, train2015/val2015 for abstract_v002).
<br>
<code>num_choices</code>: (only in MultipleChoice format) number of choices for each question (=18). For details on how the 18 choices are created, please see the <a href="http://arxiv.org/pdf/1505.00468v6.pdf" onClick="ga('send', 'event', { eventCategory: 'ICCV2015 Paper Download', eventAction: 'Paper', eventLabel: 'ICCV2015_Paper'});">VQA paper</a>.
<br>
</p>
</div>
<hr>
</div>
<div class="row">
<h1 style="font-size:30px; color:grey; font-weight: 200">Annotation Format</h1>
<div class="large-12 columns" style="text-align:center;">
<p style="font-size:15px; font-weight: 200; text-align:left;"> There is a common annotation file for the Open-Ended and Multiple-Choice tasks. The annotations are stored using the JSON file format.
<br>
<br> The annotations format has the following data structure:
<br>
</p>
<p style="font-size:15px; font-weight: 200; border-style: solid;
border-width: 1px; text-align:left; padding:5px;"><code>{<br>
"info" : info,<br>
"data_type": str,<br>
"data_subtype": str,<br>
"annotations" : [annotation],<br>
"license" : license<br>
}<br>
<br>
info {<br>
"year" : int,<br>
"version" : str,<br>
"description" : str,<br>
"contributor" : str,<br>
"url" : str,<br>
"date_created" : datetime<br>
}<br>
<br>
license{<br>
"name" : str,<br>
"url" : str<br>
}<br>
<br>
annotation{<br>
"question_id" : int,<br>
"image_id" : int,<br>
"question_type" : str,<br>
"answer_type" : str,<br>
"answers" : [answer],<br>
"multiple_choice_answer" : str<br>
}<br>
<br>
answer{<br>
"answer_id" : int,<br>
"answer" : str,<br>
"answer_confidence": str<br>
}</code></p>
<p style="font-size:15px; font-weight: 200; text-align:left;">
<code>data_type</code>: source of the images (mscoco or abstract_v002).
<br>
<code>data_subtype</code>: type of data subtype (train2014/val2014 for mscoco, train2015/val2015 for abstract_v002).
<br>
<code>question_type</code>: type of the question determined by the first few words of the question. For details, please see <a href="https://github.com/VT-vision-lab/VQA/blob/master/README.md" target="_blank">README</a>.
<br>
<code>answer_type</code>: type of the answer. Currently, "yes/no", "number", and "other".
<br>
<code>multiple_choice_answer</code>: correct multiple choice answer.
<br>
<code>answer_confidence</code>: subject's confidence in answering the question. For details, please see the <a href="http://arxiv.org/pdf/1505.00468v6.pdf">VQA paper</a>.
<br>
</p>
</div>
<hr>
</div>
<div class="row">
<h1 style="font-size:30px; color:grey; font-weight: 200">Abstract Scenes and Captions</h1>
<div class="large-12 columns" style="text-align:center;">
<p style="font-size:15px; font-weight: 200; text-align:left;"> This section provides more information regarding abstract scenes' composition (e.g., the (x,y) pixel coordinates of each clipart object, left/right facing) files and abstract captions. If you are using any data (images, questions, answers, or captions) associated with abstract scenes, please cite the <a href="http://arxiv.org/pdf/1505.00468v6.pdf">VQA paper</a>. An example BibTeX is:
<p style="font-size:15px; font-weight: 200; border-style: solid;
border-width: 1px; text-align:left; padding:5px;"><code>
@InProceedings{VQA,<br>
author = {Stanislaw Antol and Aishwarya Agrawal and Jiasen Lu and Margaret Mitchell and Dhruv Batra and C. Lawrence Zitnick and Devi Parikh},<br>
title = {VQA: Visual Question Answering},<br>
booktitle = {International Conference on Computer Vision (ICCV)},<br>
year = {2015},<br>
}
</code>
</p>
<div style="display: none;" id="hiddenText">
<p style="font-size:12px; font-weight: 200; width:150%; margin-left:-40px;">
}</p>
</div>
</p>
<p style="font-size:15px; font-weight: 200; text-align:left;">The following links contain the abstract scenes' composition files:
</p>
<div class="large-12 columns">
<ul style="font-size:15px; font-weight: 200; text-align:left;">
<li><a href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/scene_json/scene_json_abstract_v002_train2015.zip" onClick="ga('send', 'event', { eventCategory: 'Training Set Download', eventAction: 'Training Compositions', eventLabel: 'Train_Compositions_Abstract'});">train2015 Composition Files</a></li>
<li><a href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/scene_json/scene_json_abstract_v002_val2015.zip" onClick="ga('send', 'event', { eventCategory: 'Validation Set Download', eventAction: 'Validation Compositions', eventLabel: 'Val_Compositions_Abstract'});">val2015 Composition Files</a></li>
<li><a href="https://s3.amazonaws.com/cvmlp/vqa/abstract_v002/scene_json/scene_json_abstract_v002_test2015.zip" onClick="ga('send', 'event', { eventCategory: 'Test Set Download', eventAction: 'Test Compositions', eventLabel: 'Test_Compositions_Abstract'});">test2015 Composition Files</a></li>
</ul>
<br>
</div>
<p style="font-size:15px; font-weight: 200; text-align:left;">
Each of the links above contains the following:
</p>
<div class="large-12 columns">
<ul style="font-size:15px; font-weight: 200; display: inline; text-align:left;">
<li>A file of the type "abstract_v002_[datasubset]_scene_information.json" where [datasubset] is either "train2015" or "val2015" or "test2015". This file has the following data structure:</li>
<p style="font-size:15px; font-weight: 200; border-style: solid;
border-width: 1px; text-align:left; padding:5px;"><code>{<br>
"info" : info,<br>
"data_type": str,<br>
"data_subtype": str,<br>
"compositions" : [composition],<br>
"images" : [image],<br>
"license" : license<br>
}<br>
<br>
info {<br>
"year" : int,<br>
"version" : str,<br>
"description" : str,<br>
"contributor" : str,<br>
"url" : str,<br>
"date_created" : datetime<br>
}<br>
<br>
license{<br>
"name" : str,<br>
"url" : str<br>
}<br>
<br>
image{<br>
"image_id" : int,<br>
"file_name" : str,<br>
"url" : str,<br>
"height" : int,<br>
"width" : int<br>
}<br>
<br>
composition{<br>
"image_id" : int,<br>
"file_name" : str<br>
}<br>
<br>
</code></p>
<p style="font-size:15px; font-weight: 200; text-align:left;">
<code>data_type</code>: source of the images (abstract_v002).
<br>
<code>data_subtype</code>: type of data subtype (train2015/val2015/test2015).
<br> The <code>file_name</code> in <code>images</code> list contains the name of the image file for the corresponding abstract scene. These image files can be downloaded from the links provided in the "Download" section in this page.
<br> The <code>file_name</code> in <code>compositions</code> list contains the name of the scene composition file for the corresponding abstract scene (see the bullet below).
</p>
<li>A folder of the type "scene_composition_abstract_v002_[datasubset]" where [datasubset] is either "train2015" or "val2015" or "test2015". This folder contains the scene composition files for the corresponding [datasubset].</li>
<ul>
<p style="font-size:15px; font-weight: 200; text-align:left;">For more information on how to render the scenes from annotation files and to obtain API support for abstract scenes, please visit the <a href="https://github.com/VT-vision-lab/abstract_scenes_v002">GitHub repository</a>.
<br>
</p>
<p style="font-size:15px; font-weight: 200; text-align:left;">
The JSON files containing the <b>captions</b> for training and validation sets of the abstract scenes can be downloaded from the link provided in the "Download" section in this page. These files have the following data structure:
<br>
</p>
<p style="font-size:15px; font-weight: 200; border-style: solid;
border-width: 1px; text-align:left; padding:5px;"><code>{<br>
"info" : info,<br>
"task_type": str,<br>
"data_type": str,<br>
"data_subtype": str,<br>
"annotations" : [annotation],<br>
"images" : [image],<br>
"license" : license<br>
}<br>
<br>
info {<br>
"year" : int,<br>
"version" : str,<br>
"description" : str,<br>
"contributor" : str,<br>
"url" : str,<br>
"date_created" : datetime<br>
}<br>
<br>
license{<br>
"name" : str,<br>
"url" : str<br>
}<br>
<br>
image{<br>
"image_id" : int,<br>
"file_name" : str,<br>
"url" : str,<br>
"height" : int,<br>
"width" : int<br>
}<br>
<br>
annotation{<br>
"id" : int,<br>
"image_id" : int,<br>
"caption" : str<br>
}<br>
<br>
</code></p>
<p style="font-size:15px; font-weight: 200; text-align:left;">
<code>task_type</code>: Captioning.
<br>
<code>data_type</code>: dataset source of the images (abstract_v002).
<br>
<code>data_subtype</code>: type of datasubset (train2015/val2015).
<br>
</p>
</div>
<hr>
<script>
(function(i, s, o, g, r, a, m) {
i['GoogleAnalyticsObject'] = r;
i[r] = i[r] || function() {
(i[r].q = i[r].q || []).push(arguments)
}, i[r].l = 1 * new Date();
a = s.createElement(o),
m = s.getElementsByTagName(o)[0];
a.async = 1;
a.src = g;
m.parentNode.insertBefore(a, m)
})(window, document, 'script', '//www.google-analytics.com/analytics.js', 'ga');
ga('create', 'UA-63638588-1', 'auto');
ga('send', 'pageview');
</script>
<!-- jquery smooth scroll to id's -->
<script>
$(function() {
$('a[href*=#]:not([href=#])').click(function() {
if (location.pathname.replace(/^\//, '') == this.pathname.replace(/^\//, '') && location.hostname == this.hostname) {
var target = $(this.hash);
target = target.length ? target : $('[name=' + this.hash.slice(1) + ']');
if (target.length) {
$('html,body').animate({
scrollTop: target.offset().top
}, 1000);
return false;
}
}
});
});
</script>
</body>
</html>