Remote Sensing Referring Expression Understanding
Collection
REU task for RS.
•
5 items
•
Updated
Dataset Viewer
images
images listlengths 1
1
| problem
stringlengths 407
496
| answer
stringlengths 10
99
| bbox_list
listlengths 0
121
| mask
images listlengths 1
1
| category
stringclasses 20
values |
|---|---|---|---|---|---|
<image>Please carefully check the image and answer: The gray small windmill. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The gray small windmill
|
[
[
360,
382,
362,
384
],
[
360,
348,
361,
352
],
[
381,
304,
383,
320
],
[
379,
272,
380,
273
],
[
360,
272,
384,
384
]
] |
windmill
|
||
<image>Please carefully check the image and answer: A tennis court on the lower left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A tennis court on the lower left
|
[
[
223,
647,
224,
648
],
[
176,
544,
224,
648
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: A large frustum of a cone chimney. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large frustum of a cone chimney
|
[
[
395,
639,
397,
640
],
[
302,
623,
303,
624
],
[
454,
614,
456,
616
],
[
463,
607,
464,
608
],
[
471,
599,
472,
600
],
[
471,
597,
472,
598
],
[
264,
594,
265,
596
],
[
252,
575,
253,
576
],
[
493,
566,
496,
568
],
[
495,
552,
496,
553
],
[
240,
550,
243,
553
],
[
240,
548,
241,
549
],
[
240,
533,
241,
535
],
[
264,
426,
265,
427
],
[
482,
424,
483,
425
],
[
264,
424,
265,
425
],
[
308,
392,
309,
393
],
[
304,
392,
307,
396
],
[
352,
312,
353,
313
],
[
240,
304,
504,
640
]
] |
chimney
|
||
<image>Please carefully check the image and answer: The tennis court on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The tennis court on the right
|
[
[
636,
487,
637,
488
],
[
544,
383,
545,
384
],
[
544,
304,
640,
488
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: The airport in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The airport in the middle
|
[
[
432,
476,
433,
477
],
[
463,
470,
464,
471
],
[
467,
463,
468,
464
],
[
448,
449,
449,
450
],
[
452,
448,
453,
449
],
[
456,
425,
458,
426
],
[
512,
321,
513,
322
],
[
535,
317,
536,
318
],
[
400,
312,
544,
544
]
] |
airport
|
||
<image>Please carefully check the image and answer: The expressway toll station in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The expressway toll station in the middle
|
[
[
472,
367,
473,
368
],
[
511,
353,
512,
354
],
[
511,
315,
512,
316
],
[
508,
312,
509,
313
],
[
483,
208,
484,
209
],
[
440,
208,
512,
368
]
] |
Expressway-toll-station
|
||
<image>Please carefully check the image and answer: The basketball court is on the right of the tennis court in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The basketball court is on the right of the tennis court in the middle
|
[
[
503,
423,
504,
424
],
[
424,
375,
425,
377
],
[
511,
326,
512,
327
],
[
501,
257,
504,
261
],
[
495,
240,
496,
241
],
[
419,
240,
421,
241
],
[
400,
240,
512,
424
]
] |
basketballcourt
|
||
<image>Please carefully check the image and answer: A oval ground track field. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A oval ground track field
|
[
[
406,
750,
407,
752
],
[
240,
719,
241,
720
],
[
495,
655,
496,
656
],
[
200,
654,
203,
656
],
[
200,
648,
201,
650
],
[
200,
637,
201,
639
],
[
503,
600,
504,
601
],
[
502,
561,
504,
564
],
[
503,
495,
504,
496
],
[
503,
492,
504,
494
],
[
503,
455,
504,
456
],
[
495,
236,
496,
239
],
[
495,
233,
496,
235
],
[
494,
198,
496,
202
],
[
495,
196,
496,
197
],
[
216,
128,
217,
131
],
[
232,
112,
233,
113
],
[
431,
100,
432,
101
],
[
273,
88,
274,
89
],
[
399,
81,
400,
83
],
[
184,
72,
504,
760
]
] |
groundtrackfield
|
||
<image>Please carefully check the image and answer: The large expressway toll station. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The large expressway toll station
|
[
[
599,
452,
600,
455
],
[
408,
401,
409,
402
],
[
400,
392,
600,
496
]
] |
Expressway-toll-station
|
||
<image>Please carefully check the image and answer: A windmill on the lower left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A windmill on the lower left
|
[
[
208,
692,
210,
694
],
[
195,
656,
197,
657
],
[
160,
624,
224,
695
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The vehicle is a little smaller than the vehicle in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The vehicle is a little smaller than the vehicle in the middle
|
[
[
600,
386,
601,
387
],
[
671,
368,
672,
369
],
[
600,
360,
672,
392
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A slender green golf field in the top. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A slender green golf field in the top
|
[
[
263,
399,
264,
400
],
[
216,
398,
218,
400
],
[
487,
382,
490,
384
],
[
216,
376,
217,
378
],
[
652,
365,
656,
368
],
[
671,
349,
672,
352
],
[
216,
343,
217,
344
],
[
687,
327,
688,
328
],
[
680,
272,
681,
274
],
[
703,
241,
704,
242
],
[
703,
239,
704,
240
],
[
700,
224,
704,
230
],
[
536,
224,
537,
225
],
[
343,
216,
344,
217
],
[
610,
208,
613,
210
],
[
408,
208,
409,
209
],
[
216,
208,
712,
400
]
] |
golffield
|
||
<image>Please carefully check the image and answer: The frustum of a cone blue and gray huge chimney. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The frustum of a cone blue and gray huge chimney
|
[
[
335,
655,
336,
656
],
[
342,
647,
343,
648
],
[
346,
639,
347,
640
],
[
104,
590,
105,
592
],
[
104,
577,
105,
578
],
[
104,
575,
105,
576
],
[
123,
488,
124,
489
],
[
551,
487,
552,
489
],
[
559,
474,
560,
475
],
[
575,
454,
576,
456
],
[
583,
437,
584,
438
],
[
176,
415,
177,
416
],
[
599,
368,
600,
369
],
[
176,
368,
178,
372
],
[
599,
320,
600,
321
],
[
200,
240,
202,
241
],
[
208,
232,
212,
235
],
[
224,
214,
226,
215
],
[
248,
184,
249,
185
],
[
510,
168,
512,
170
],
[
428,
136,
429,
137
],
[
380,
136,
381,
137
],
[
369,
136,
370,
137
],
[
104,
136,
600,
696
]
] |
chimney
|
||
<image>Please carefully check the image and answer: A large airport in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large airport in the middle
|
[
[
0,
463,
1,
464
],
[
10,
408,
12,
409
],
[
231,
393,
232,
395
],
[
232,
340,
233,
341
],
[
280,
320,
285,
322
],
[
560,
287,
561,
288
],
[
584,
276,
589,
280
],
[
630,
263,
632,
264
],
[
464,
260,
465,
261
],
[
571,
240,
572,
241
],
[
549,
232,
550,
233
],
[
733,
222,
736,
224
],
[
709,
221,
723,
224
],
[
592,
221,
600,
224
],
[
741,
214,
744,
216
],
[
692,
200,
697,
203
],
[
728,
190,
729,
191
],
[
759,
177,
760,
178
],
[
771,
176,
772,
178
],
[
0,
176,
776,
464
]
] |
airport
|
||
<image>Please carefully check the image and answer: A vehicle is on the right of the tiny expressway toll station at the bottom. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle is on the right of the tiny expressway toll station at the bottom
|
[
[
656,
608,
680,
624
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The vehicle is on the upper left of the vehicle on the lower left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The vehicle is on the upper left of the vehicle on the lower left
|
[
[
39,
154,
40,
155
],
[
16,
128,
40,
160
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A vehicle on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle on the right
|
[
[
576,
562,
577,
564
],
[
576,
560,
577,
561
],
[
607,
512,
608,
513
],
[
568,
506,
569,
512
],
[
568,
496,
616,
584
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The expressway service area on the top. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The expressway service area on the top
|
[
[
647,
359,
649,
360
],
[
264,
318,
265,
319
],
[
687,
303,
688,
304
],
[
304,
292,
306,
294
],
[
685,
241,
686,
242
],
[
384,
236,
385,
237
],
[
429,
208,
430,
209
],
[
448,
187,
449,
188
],
[
496,
138,
497,
139
],
[
678,
136,
679,
137
],
[
512,
128,
513,
129
],
[
678,
126,
680,
128
],
[
523,
120,
524,
121
],
[
679,
118,
680,
119
],
[
528,
112,
529,
113
],
[
536,
96,
540,
98
],
[
585,
66,
587,
67
],
[
594,
64,
595,
65
],
[
264,
64,
688,
368
],
[
584,
64,
587,
65
]
] |
Expressway-Service-area
|
||
<image>Please carefully check the image and answer: A oval ground track field on the left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A oval ground track field on the left
|
[
[
199,
787,
200,
788
],
[
137,
783,
138,
784
],
[
247,
781,
248,
782
],
[
137,
781,
138,
782
],
[
262,
770,
264,
772
],
[
283,
758,
287,
760
],
[
303,
749,
304,
750
],
[
64,
735,
65,
736
],
[
64,
733,
65,
734
],
[
335,
719,
336,
720
],
[
335,
708,
336,
709
],
[
16,
654,
17,
655
],
[
16,
652,
17,
653
],
[
357,
633,
358,
634
],
[
357,
600,
360,
633
],
[
16,
484,
17,
485
],
[
16,
310,
17,
311
],
[
16,
308,
17,
309
],
[
21,
136,
22,
137
],
[
343,
113,
344,
115
],
[
335,
100,
336,
101
],
[
335,
97,
336,
99
],
[
47,
68,
48,
69
],
[
80,
56,
83,
58
],
[
96,
40,
97,
41
],
[
112,
32,
113,
33
],
[
247,
24,
248,
25
],
[
212,
24,
213,
25
],
[
139,
24,
140,
25
],
[
16,
24,
360,
792
]
] |
groundtrackfield
|
||
<image>Please carefully check the image and answer: The small expressway toll station. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The small expressway toll station
|
[
[
447,
456,
448,
457
],
[
373,
456,
374,
458
],
[
368,
360,
472,
496
]
] |
Expressway-toll-station
|
||
<image>Please carefully check the image and answer: A ship on the left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A ship on the left
|
[
[
237,
351,
238,
352
],
[
216,
336,
240,
368
],
[
216,
336,
217,
339
]
] |
ship
|
||
<image>Please carefully check the image and answer: A gray overpass in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A gray overpass in the middle
|
[
[
471,
583,
472,
584
],
[
469,
583,
470,
584
],
[
471,
581,
472,
582
],
[
486,
567,
487,
568
],
[
328,
549,
329,
550
],
[
512,
525,
513,
526
],
[
368,
500,
369,
502
],
[
376,
490,
377,
491
],
[
376,
488,
377,
489
],
[
384,
481,
388,
483
],
[
384,
480,
385,
481
],
[
392,
473,
393,
474
],
[
607,
423,
608,
424
],
[
607,
421,
608,
422
],
[
456,
394,
457,
395
],
[
457,
392,
458,
393
],
[
464,
384,
465,
385
],
[
634,
383,
636,
384
],
[
655,
359,
656,
360
],
[
655,
357,
656,
358
],
[
488,
353,
489,
354
],
[
671,
334,
672,
335
],
[
520,
313,
521,
314
],
[
528,
296,
530,
297
],
[
552,
278,
553,
279
],
[
328,
224,
688,
624
],
[
584,
224,
585,
225
]
] |
overpass
|
||
<image>Please carefully check the image and answer: A tiny vehicle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A tiny vehicle
|
[
[
560,
279,
561,
280
],
[
552,
264,
553,
265
],
[
599,
256,
600,
257
],
[
552,
248,
600,
280
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The gray large overpass in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The gray large overpass in the middle
|
[
[
480,
631,
481,
632
],
[
280,
200,
488,
632
]
] |
overpass
|
||
<image>Please carefully check the image and answer: The tiny ship. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The tiny ship
|
[
[
408,
507,
410,
508
],
[
479,
450,
480,
451
],
[
408,
448,
480,
520
]
] |
ship
|
||
<image>Please carefully check the image and answer: A vehicle on the upper right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle on the upper right
|
[
[
671,
127,
672,
128
],
[
640,
127,
641,
128
],
[
640,
103,
641,
104
],
[
640,
96,
688,
128
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The oval ground track field. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The oval ground track field
|
[
[
383,
746,
384,
747
],
[
383,
739,
384,
740
],
[
248,
733,
249,
736
],
[
240,
643,
241,
644
],
[
240,
636,
241,
637
],
[
240,
575,
241,
577
],
[
240,
544,
241,
545
],
[
334,
497,
336,
499
],
[
314,
496,
315,
497
],
[
240,
496,
384,
776
]
] |
groundtrackfield
|
||
<image>Please carefully check the image and answer: A airport in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A airport in the middle
|
[
[
479,
595,
480,
596
],
[
440,
532,
441,
534
],
[
455,
443,
456,
444
],
[
452,
440,
453,
441
],
[
408,
333,
410,
334
],
[
408,
328,
409,
329
],
[
400,
287,
401,
288
],
[
400,
259,
401,
260
],
[
415,
242,
416,
243
],
[
395,
239,
396,
240
],
[
392,
226,
393,
228
],
[
392,
224,
393,
225
],
[
392,
211,
393,
212
],
[
384,
192,
385,
193
],
[
384,
144,
488,
648
]
] |
airport
|
||
<image>Please carefully check the image and answer: A red vehicle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A red vehicle
|
[
[
200,
46,
202,
48
],
[
207,
43,
208,
44
],
[
168,
16,
208,
48
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A slender bridge. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A slender bridge
|
[
[
438,
519,
439,
520
],
[
439,
517,
441,
519
],
[
437,
517,
438,
518
],
[
408,
488,
409,
489
],
[
460,
408,
461,
409
],
[
447,
336,
448,
337
],
[
436,
296,
437,
297
],
[
368,
272,
369,
273
],
[
431,
262,
432,
264
],
[
429,
256,
432,
258
],
[
415,
215,
416,
216
],
[
360,
212,
361,
213
],
[
360,
208,
361,
209
],
[
360,
200,
472,
520
]
] |
bridge
|
||
<image>Please carefully check the image and answer: A large windmill. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large windmill
|
[
[
353,
519,
354,
520
],
[
404,
509,
408,
512
],
[
408,
455,
409,
456
],
[
352,
408,
448,
520
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The expressway service area on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The expressway service area on the right
|
[
[
564,
671,
565,
672
],
[
552,
668,
553,
669
],
[
542,
663,
543,
664
],
[
537,
663,
538,
664
],
[
500,
655,
501,
656
],
[
412,
630,
415,
632
],
[
368,
623,
369,
624
],
[
651,
615,
652,
616
],
[
360,
610,
361,
611
],
[
368,
567,
369,
568
],
[
368,
560,
369,
561
],
[
384,
534,
386,
535
],
[
711,
530,
712,
531
],
[
711,
526,
712,
527
],
[
416,
446,
418,
449
],
[
416,
441,
417,
445
],
[
416,
432,
420,
435
],
[
416,
432,
417,
433
],
[
758,
422,
760,
425
],
[
752,
416,
754,
418
],
[
719,
400,
720,
401
],
[
687,
336,
688,
337
],
[
522,
282,
524,
283
],
[
605,
256,
608,
258
],
[
536,
249,
538,
251
],
[
536,
247,
537,
248
],
[
537,
240,
538,
241
],
[
360,
224,
760,
680
]
] |
Expressway-Service-area
|
||
<image>Please carefully check the image and answer: The gray vehicle on the lower left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The gray vehicle on the lower left
|
[
[
128,
751,
129,
752
],
[
128,
728,
168,
752
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A airplane in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A airplane in the middle
|
[
[
295,
437,
296,
438
],
[
272,
422,
273,
423
],
[
306,
384,
307,
385
],
[
326,
376,
330,
378
],
[
310,
375,
311,
376
],
[
311,
373,
312,
374
],
[
289,
344,
290,
345
],
[
313,
328,
316,
330
],
[
272,
320,
344,
464
]
] |
airplane
|
||
<image>Please carefully check the image and answer: The small bridge. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The small bridge
|
[
[
343,
436,
344,
438
],
[
442,
431,
443,
432
],
[
296,
388,
297,
389
],
[
296,
288,
656,
448
]
] |
bridge
|
||
<image>Please carefully check the image and answer: A storage tank is on the lower right of the airplane in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A storage tank is on the lower right of the airplane in the middle
|
[
[
710,
438,
712,
440
],
[
696,
408,
736,
440
]
] |
storagetank
|
||
<image>Please carefully check the image and answer: The tennis court on the left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The tennis court on the left
|
[
[
152,
447,
153,
448
],
[
175,
444,
176,
445
],
[
131,
436,
135,
440
],
[
130,
434,
131,
435
],
[
115,
430,
116,
432
],
[
183,
420,
184,
421
],
[
183,
412,
184,
413
],
[
196,
383,
197,
384
],
[
96,
375,
97,
376
],
[
223,
310,
224,
312
],
[
196,
256,
197,
257
],
[
144,
243,
145,
244
],
[
88,
240,
232,
448
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: A basketball court in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A basketball court in the middle
|
[
[
464,
351,
465,
352
],
[
535,
272,
536,
273
],
[
528,
272,
529,
273
],
[
490,
272,
494,
273
],
[
464,
272,
536,
400
]
] |
basketballcourt
|
||
<image>Please carefully check the image and answer: The windmill at the bottom. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The windmill at the bottom
|
[
[
511,
686,
512,
687
],
[
508,
668,
512,
672
],
[
534,
648,
536,
649
],
[
523,
648,
534,
653
],
[
488,
648,
491,
650
],
[
464,
640,
536,
696
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The windmill on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The windmill on the right
|
[
[
695,
466,
696,
467
],
[
694,
440,
696,
441
],
[
680,
440,
696,
472
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The orange and gray tiny overpass. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The orange and gray tiny overpass
|
[
[
456,
695,
457,
696
],
[
448,
687,
449,
688
],
[
495,
616,
496,
617
],
[
487,
608,
488,
609
],
[
448,
608,
496,
696
]
] |
overpass
|
||
<image>Please carefully check the image and answer: A ship at the bottom. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A ship at the bottom
|
[
[
400,
725,
401,
726
],
[
448,
720,
449,
721
],
[
407,
720,
408,
721
],
[
400,
720,
496,
752
]
] |
ship
|
||
<image>Please carefully check the image and answer: The baseball field on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The baseball field on the right
|
[
[
743,
404,
744,
405
],
[
758,
367,
760,
368
],
[
617,
351,
618,
352
],
[
608,
342,
609,
345
],
[
735,
304,
736,
305
],
[
608,
288,
760,
440
]
] |
baseballfield
|
||
<image>Please carefully check the image and answer: A storage tank on the lower left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A storage tank on the lower left
|
[
[
199,
620,
200,
621
],
[
207,
614,
208,
615
],
[
136,
607,
137,
608
],
[
128,
594,
129,
596
],
[
215,
582,
216,
583
],
[
128,
578,
129,
585
],
[
191,
550,
192,
551
],
[
128,
544,
216,
624
]
] |
storagetank
|
||
<image>Please carefully check the image and answer: A chimney at the bottom. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A chimney at the bottom
|
[
[
552,
560,
554,
562
],
[
504,
560,
576,
592
]
] |
chimney
|
||
<image>Please carefully check the image and answer: A blue and gray train station in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A blue and gray train station in the middle
|
[
[
479,
495,
480,
496
],
[
432,
495,
433,
496
],
[
404,
495,
405,
496
],
[
551,
455,
552,
456
],
[
279,
455,
280,
456
],
[
234,
455,
236,
456
],
[
247,
451,
248,
452
],
[
551,
448,
552,
449
],
[
464,
384,
504,
385
],
[
176,
384,
552,
496
]
] |
trainstation
|
||
<image>Please carefully check the image and answer: A blue small expressway toll station. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A blue small expressway toll station
|
[
[
366,
470,
368,
472
],
[
288,
469,
289,
472
],
[
367,
454,
368,
456
],
[
288,
435,
289,
438
],
[
280,
397,
281,
399
],
[
280,
312,
368,
480
]
] |
Expressway-toll-station
|
||
<image>Please carefully check the image and answer: The vehicle is similar in size to the vehicle on the lower left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The vehicle is similar in size to the vehicle on the lower left
|
[
[
208,
504,
232,
552
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The gray and black slender small ship. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The gray and black slender small ship
|
[
[
301,
143,
302,
144
],
[
226,
135,
227,
136
],
[
224,
135,
225,
136
],
[
224,
133,
225,
134
],
[
224,
88,
225,
89
],
[
310,
72,
311,
73
],
[
216,
72,
440,
152
]
] |
ship
|
||
<image>Please carefully check the image and answer: A chimney on the upper left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A chimney on the upper left
|
[
[
38,
8,
39,
9
],
[
31,
4,
32,
5
],
[
0,
0,
96,
56
]
] |
chimney
|
||
<image>Please carefully check the image and answer: The airplane on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The airplane on the right
|
[
[
674,
468,
677,
472
],
[
677,
466,
680,
472
],
[
624,
448,
628,
456
],
[
632,
439,
633,
442
],
[
653,
416,
656,
422
],
[
616,
400,
688,
472
],
[
616,
400,
617,
401
]
] |
airplane
|
||
<image>Please carefully check the image and answer: The overpass is on the lower right of the small gray vehicle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The overpass is on the lower right of the small gray vehicle
|
[
[
488,
607,
489,
609
],
[
488,
605,
490,
606
],
[
400,
495,
402,
505
],
[
519,
440,
520,
441
],
[
519,
399,
520,
400
],
[
496,
312,
497,
313
],
[
400,
312,
520,
624
]
] |
overpass
|
||
<image>Please carefully check the image and answer: A vehicle is above the blue and gray airplane. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle is above the blue and gray airplane
|
[
[
367,
271,
368,
272
],
[
431,
263,
432,
264
],
[
333,
263,
334,
264
],
[
328,
240,
432,
272
],
[
328,
240,
329,
241
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A gray and black ship on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A gray and black ship on the right
|
[
[
433,
499,
440,
504
],
[
384,
497,
385,
498
],
[
398,
400,
399,
401
],
[
663,
399,
664,
400
],
[
450,
384,
451,
386
],
[
734,
358,
736,
360
],
[
536,
328,
538,
331
],
[
775,
322,
776,
323
],
[
783,
309,
784,
310
],
[
595,
304,
596,
306
],
[
783,
267,
784,
268
],
[
352,
248,
784,
520
]
] |
ship
|
||
<image>Please carefully check the image and answer: The golf field at the bottom. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The golf field at the bottom
|
[
[
499,
735,
500,
736
],
[
492,
735,
493,
736
],
[
523,
734,
524,
736
],
[
494,
734,
498,
736
],
[
424,
723,
425,
724
],
[
567,
707,
568,
708
],
[
424,
528,
624,
736
],
[
608,
528,
609,
529
]
] |
golffield
|
||
<image>Please carefully check the image and answer: The tennis court is on the right of the large bridge in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The tennis court is on the right of the large bridge in the middle
|
[
[
650,
439,
652,
440
],
[
663,
420,
664,
421
],
[
671,
407,
672,
408
],
[
687,
342,
688,
344
],
[
687,
336,
688,
338
],
[
683,
336,
686,
344
],
[
632,
324,
633,
325
],
[
608,
320,
682,
440
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: A gray large stadium. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A gray large stadium
|
[
[
566,
543,
567,
544
],
[
470,
542,
472,
544
],
[
449,
535,
450,
536
],
[
573,
534,
576,
536
],
[
400,
503,
401,
505
],
[
599,
502,
600,
503
],
[
607,
490,
608,
491
],
[
384,
390,
385,
391
],
[
384,
368,
387,
376
],
[
384,
335,
385,
343
],
[
615,
252,
616,
253
],
[
392,
244,
393,
245
],
[
615,
243,
616,
244
],
[
400,
219,
401,
220
],
[
591,
202,
592,
203
],
[
591,
200,
592,
201
],
[
470,
176,
471,
177
],
[
384,
176,
616,
544
]
] |
stadium
|
||
<image>Please carefully check the image and answer: A ground track field on the lower right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A ground track field on the lower right
|
[
[
736,
528,
737,
529
],
[
728,
528,
752,
576
]
] |
groundtrackfield
|
||
<image>Please carefully check the image and answer: A storage tank in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A storage tank in the middle
|
[
[
246,
462,
248,
464
],
[
295,
445,
296,
448
],
[
208,
435,
209,
436
],
[
295,
408,
296,
409
],
[
208,
400,
209,
403
],
[
287,
395,
288,
396
],
[
287,
393,
288,
394
],
[
208,
384,
296,
464
]
] |
storagetank
|
||
<image>Please carefully check the image and answer: The white and gray expressway toll station in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The white and gray expressway toll station in the middle
|
[
[
367,
519,
368,
520
],
[
430,
462,
432,
464
],
[
431,
458,
432,
459
],
[
431,
441,
432,
442
],
[
304,
400,
432,
536
]
] |
Expressway-toll-station
|
||
<image>Please carefully check the image and answer: A large dam. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large dam
|
[
[
174,
510,
176,
512
],
[
385,
503,
386,
504
],
[
255,
503,
256,
504
],
[
213,
503,
214,
504
],
[
480,
495,
482,
496
],
[
494,
494,
496,
495
],
[
482,
491,
493,
496
],
[
653,
486,
656,
488
],
[
648,
486,
651,
488
],
[
91,
480,
92,
482
],
[
539,
479,
540,
480
],
[
605,
477,
607,
480
],
[
128,
472,
129,
473
],
[
223,
442,
224,
443
],
[
223,
440,
224,
441
],
[
680,
432,
681,
433
],
[
330,
432,
332,
433
],
[
444,
424,
445,
425
],
[
560,
412,
561,
413
],
[
566,
408,
567,
409
],
[
72,
392,
696,
520
]
] |
dam
|
||
<image>Please carefully check the image and answer: A windmill on the lower right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A windmill on the lower right
|
[
[
760,
627,
762,
628
],
[
765,
616,
766,
617
],
[
719,
584,
720,
585
],
[
669,
576,
670,
577
],
[
656,
576,
776,
632
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The small train station. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The small train station
|
[
[
591,
255,
592,
256
],
[
512,
255,
513,
256
],
[
528,
234,
529,
235
],
[
528,
232,
529,
233
],
[
536,
226,
537,
227
],
[
536,
224,
537,
225
],
[
610,
223,
611,
224
],
[
560,
200,
564,
203
],
[
584,
184,
588,
188
],
[
512,
160,
640,
280
]
] |
trainstation
|
||
<image>Please carefully check the image and answer: A large cyan and gray stadium. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large cyan and gray stadium
|
[
[
480,
527,
481,
528
],
[
535,
471,
536,
472
],
[
540,
463,
541,
464
],
[
313,
461,
315,
464
],
[
296,
352,
297,
353
],
[
567,
351,
568,
352
],
[
567,
321,
568,
323
],
[
567,
318,
568,
320
],
[
304,
286,
305,
287
],
[
567,
265,
568,
266
],
[
304,
245,
305,
246
],
[
336,
165,
337,
166
],
[
535,
157,
536,
158
],
[
527,
146,
528,
147
],
[
527,
144,
528,
145
],
[
369,
128,
371,
129
],
[
383,
120,
384,
121
],
[
459,
112,
460,
113
],
[
297,
112,
568,
528
]
] |
stadium
|
||
<image>Please carefully check the image and answer: A small bridge. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A small bridge
|
[
[
183,
280,
184,
281
],
[
142,
279,
143,
280
],
[
179,
262,
180,
264
],
[
88,
254,
90,
256
],
[
31,
138,
32,
139
],
[
31,
136,
32,
137
],
[
23,
130,
24,
131
],
[
23,
128,
24,
129
],
[
14,
120,
16,
123
],
[
7,
115,
8,
116
],
[
0,
112,
184,
296
],
[
0,
112,
3,
114
]
] |
bridge
|
||
<image>Please carefully check the image and answer: A train station in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A train station in the middle
|
[
[
224,
528,
227,
537
],
[
224,
520,
225,
528
],
[
226,
496,
227,
520
],
[
224,
488,
225,
496
],
[
295,
479,
296,
480
],
[
224,
470,
226,
472
],
[
224,
456,
226,
459
],
[
224,
392,
225,
400
],
[
224,
382,
225,
383
],
[
224,
358,
226,
360
],
[
295,
356,
296,
359
],
[
224,
356,
225,
357
],
[
294,
352,
296,
354
],
[
295,
304,
296,
305
],
[
224,
240,
225,
241
],
[
224,
232,
296,
552
]
] |
trainstation
|
||
<image>Please carefully check the image and answer: A tennis court in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A tennis court in the middle
|
[
[
520,
508,
521,
509
],
[
504,
487,
505,
488
],
[
506,
464,
508,
465
],
[
504,
464,
505,
465
],
[
504,
448,
592,
552
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: The vehicle is above the small expressway toll station on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The vehicle is above the small expressway toll station on the right
|
[
[
680,
303,
681,
304
],
[
680,
288,
704,
304
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A tennis court is above the tennis court on the lower right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A tennis court is above the tennis court on the lower right
|
[
[
464,
506,
465,
507
],
[
456,
501,
457,
503
],
[
534,
500,
536,
501
],
[
440,
477,
441,
478
],
[
574,
471,
575,
472
],
[
432,
468,
433,
469
],
[
440,
440,
444,
442
],
[
512,
386,
513,
387
],
[
512,
384,
513,
385
],
[
639,
383,
640,
385
],
[
612,
352,
614,
353
],
[
432,
328,
648,
528
],
[
584,
328,
585,
329
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: The airplane at the bottom. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The airplane at the bottom
|
[
[
390,
688,
392,
689
],
[
344,
680,
392,
712
]
] |
airplane
|
||
<image>Please carefully check the image and answer: A small overpass. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A small overpass
|
[
[
336,
407,
337,
408
],
[
336,
398,
337,
399
],
[
367,
380,
368,
382
],
[
328,
249,
329,
250
],
[
328,
247,
329,
248
],
[
367,
208,
368,
211
],
[
328,
208,
368,
424
]
] |
overpass
|
||
<image>Please carefully check the image and answer: The small overpass. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The small overpass
|
[
[
7,
166,
8,
167
],
[
39,
143,
40,
144
],
[
37,
143,
38,
144
],
[
39,
141,
40,
142
],
[
47,
135,
48,
136
],
[
48,
90,
49,
91
],
[
48,
88,
49,
89
],
[
56,
80,
60,
84
],
[
72,
67,
74,
68
],
[
143,
44,
144,
45
],
[
104,
43,
105,
44
],
[
183,
8,
184,
9
],
[
188,
2,
192,
8
],
[
0,
0,
192,
168
]
] |
overpass
|
||
<image>Please carefully check the image and answer: A large cyan and gray airport. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large cyan and gray airport
|
[
[
536,
575,
539,
577
],
[
165,
575,
168,
576
],
[
159,
575,
163,
576
],
[
176,
572,
181,
576
],
[
531,
567,
532,
568
],
[
152,
563,
153,
564
],
[
528,
557,
529,
558
],
[
239,
546,
240,
547
],
[
527,
520,
528,
521
],
[
514,
496,
517,
497
],
[
348,
486,
349,
488
],
[
488,
477,
489,
479
],
[
503,
476,
504,
478
],
[
464,
440,
465,
441
],
[
360,
432,
361,
433
],
[
465,
376,
466,
377
],
[
472,
368,
476,
371
],
[
424,
355,
425,
356
],
[
441,
352,
442,
353
],
[
416,
329,
418,
330
],
[
430,
328,
432,
329
],
[
419,
328,
430,
333
],
[
152,
312,
656,
640
]
] |
airport
|
||
<image>Please carefully check the image and answer: The gray vehicle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The gray vehicle
|
[
[
271,
422,
272,
423
],
[
214,
408,
216,
410
],
[
200,
408,
272,
440
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The small slender overpass. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The small slender overpass
|
[
[
376,
415,
377,
416
],
[
368,
400,
369,
401
],
[
368,
360,
369,
361
],
[
437,
335,
439,
336
],
[
439,
303,
440,
304
],
[
439,
279,
440,
280
],
[
447,
166,
448,
168
],
[
446,
105,
448,
106
],
[
368,
96,
456,
416
],
[
392,
96,
393,
104
]
] |
overpass
|
||
<image>Please carefully check the image and answer: A windmill on the lower right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A windmill on the lower right
|
[
[
600,
711,
603,
712
],
[
600,
707,
608,
710
],
[
544,
639,
545,
640
],
[
575,
632,
576,
633
],
[
569,
632,
570,
633
],
[
536,
619,
539,
622
],
[
536,
616,
608,
706
],
[
536,
616,
538,
618
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The golf field in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The golf field in the middle
|
[
[
378,
639,
379,
640
],
[
352,
638,
354,
639
],
[
391,
627,
392,
628
],
[
312,
607,
313,
608
],
[
420,
592,
421,
593
],
[
312,
587,
313,
588
],
[
448,
511,
449,
512
],
[
399,
480,
400,
481
],
[
495,
477,
496,
478
],
[
296,
469,
297,
470
],
[
296,
456,
297,
457
],
[
453,
455,
454,
456
],
[
573,
454,
574,
455
],
[
583,
453,
584,
454
],
[
560,
443,
561,
444
],
[
305,
434,
306,
435
],
[
305,
432,
306,
433
],
[
407,
415,
408,
416
],
[
575,
408,
576,
409
],
[
503,
392,
504,
393
],
[
320,
366,
322,
368
],
[
574,
359,
575,
360
],
[
304,
295,
305,
296
],
[
602,
263,
603,
264
],
[
293,
240,
294,
241
],
[
299,
232,
300,
233
],
[
304,
224,
305,
225
],
[
320,
222,
322,
223
],
[
320,
216,
322,
219
],
[
605,
208,
608,
212
],
[
399,
204,
400,
205
],
[
416,
200,
417,
201
],
[
383,
200,
384,
201
],
[
433,
194,
434,
195
],
[
433,
192,
434,
193
],
[
530,
176,
533,
177
],
[
280,
176,
608,
640
]
] |
golffield
|
||
<image>Please carefully check the image and answer: A large expressway service area in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A large expressway service area in the middle
|
[
[
544,
679,
545,
680
],
[
536,
671,
537,
672
],
[
488,
655,
489,
656
],
[
471,
655,
472,
656
],
[
471,
648,
472,
649
],
[
408,
623,
409,
624
],
[
397,
614,
399,
616
],
[
575,
606,
576,
607
],
[
328,
573,
329,
574
],
[
288,
550,
292,
552
],
[
272,
543,
273,
544
],
[
248,
527,
249,
528
],
[
583,
488,
584,
489
],
[
160,
479,
161,
480
],
[
152,
469,
153,
470
],
[
128,
463,
129,
464
],
[
113,
455,
114,
456
],
[
113,
453,
114,
454
],
[
559,
448,
560,
449
],
[
527,
432,
528,
433
],
[
511,
424,
512,
425
],
[
496,
424,
497,
425
],
[
490,
424,
492,
426
],
[
104,
416,
105,
417
],
[
117,
408,
119,
410
],
[
152,
384,
156,
388
],
[
407,
378,
408,
379
],
[
396,
368,
400,
370
],
[
351,
368,
352,
369
],
[
184,
360,
185,
361
],
[
200,
354,
201,
355
],
[
295,
338,
296,
339
],
[
88,
336,
600,
680
]
] |
Expressway-Service-area
|
||
<image>Please carefully check the image and answer: The airport in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The airport in the middle
|
[
[
127,
631,
128,
632
],
[
135,
623,
136,
624
],
[
101,
623,
102,
624
],
[
93,
615,
94,
616
],
[
159,
607,
160,
608
],
[
204,
591,
205,
592
],
[
196,
591,
197,
592
],
[
191,
591,
192,
592
],
[
215,
589,
216,
590
],
[
228,
582,
232,
584
],
[
269,
567,
270,
568
],
[
80,
560,
82,
564
],
[
88,
552,
92,
554
],
[
311,
549,
312,
550
],
[
367,
525,
368,
526
],
[
171,
520,
172,
521
],
[
160,
520,
164,
522
],
[
383,
519,
384,
520
],
[
372,
519,
373,
520
],
[
187,
512,
188,
513
],
[
399,
511,
400,
512
],
[
240,
490,
241,
491
],
[
265,
482,
266,
483
],
[
265,
480,
266,
481
],
[
471,
479,
473,
480
],
[
280,
472,
282,
474
],
[
494,
466,
496,
468
],
[
296,
464,
297,
465
],
[
320,
458,
321,
459
],
[
349,
448,
350,
449
],
[
360,
440,
362,
442
],
[
400,
424,
401,
425
],
[
615,
421,
616,
422
],
[
483,
408,
484,
409
],
[
480,
408,
482,
409
],
[
655,
405,
656,
406
],
[
671,
399,
672,
400
],
[
668,
397,
672,
400
],
[
727,
373,
728,
374
],
[
751,
365,
752,
366
],
[
767,
359,
768,
360
],
[
780,
350,
784,
352
],
[
567,
306,
568,
307
],
[
688,
288,
689,
289
],
[
448,
285,
449,
286
],
[
720,
284,
721,
285
],
[
527,
274,
528,
275
],
[
762,
272,
763,
273
],
[
397,
239,
398,
240
],
[
447,
208,
448,
209
],
[
439,
200,
440,
201
],
[
423,
192,
424,
193
],
[
415,
184,
416,
185
],
[
314,
183,
315,
184
],
[
290,
167,
291,
168
],
[
380,
160,
384,
162
],
[
256,
139,
259,
144
],
[
248,
135,
249,
136
],
[
284,
88,
288,
90
],
[
160,
62,
164,
64
],
[
157,
55,
158,
56
],
[
80,
0,
784,
632
]
] |
airport
|
||
<image>Please carefully check the image and answer: A orange and red airport. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A orange and red airport
|
[
[
440,
783,
441,
784
],
[
440,
773,
441,
774
],
[
687,
772,
688,
773
],
[
437,
767,
438,
768
],
[
416,
743,
417,
744
],
[
405,
711,
406,
712
],
[
655,
704,
656,
705
],
[
634,
680,
635,
681
],
[
384,
675,
385,
676
],
[
381,
663,
382,
664
],
[
360,
631,
361,
632
],
[
599,
615,
600,
617
],
[
352,
610,
353,
611
],
[
341,
591,
342,
592
],
[
586,
576,
587,
577
],
[
333,
575,
334,
576
],
[
583,
550,
584,
551
],
[
583,
547,
584,
549
],
[
575,
513,
576,
514
],
[
573,
512,
574,
513
],
[
293,
503,
294,
504
],
[
288,
490,
289,
491
],
[
272,
466,
273,
467
],
[
248,
424,
249,
425
],
[
522,
416,
523,
417
],
[
519,
412,
520,
413
],
[
519,
409,
520,
411
],
[
503,
387,
504,
388
],
[
224,
383,
225,
384
],
[
495,
360,
496,
361
],
[
471,
320,
472,
321
],
[
176,
297,
177,
298
],
[
431,
264,
432,
265
],
[
152,
247,
153,
248
],
[
375,
184,
376,
185
],
[
362,
168,
363,
169
],
[
359,
152,
360,
153
],
[
351,
150,
352,
151
],
[
341,
128,
344,
130
],
[
104,
124,
105,
125
],
[
335,
121,
336,
122
],
[
105,
117,
107,
118
],
[
334,
114,
335,
115
],
[
334,
112,
335,
113
],
[
208,
52,
209,
53
],
[
199,
52,
200,
53
],
[
184,
50,
185,
51
],
[
263,
40,
264,
41
],
[
104,
40,
704,
800
]
] |
airport
|
||
<image>Please carefully check the image and answer: A golf field on the left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A golf field on the left
|
[
[
312,
783,
313,
784
],
[
123,
783,
124,
784
],
[
343,
704,
344,
706
],
[
96,
680,
97,
681
],
[
343,
660,
344,
661
],
[
318,
647,
319,
648
],
[
334,
646,
336,
648
],
[
104,
632,
105,
633
],
[
351,
591,
352,
592
],
[
351,
534,
352,
535
],
[
350,
530,
352,
533
],
[
351,
466,
352,
468
],
[
351,
464,
352,
465
],
[
112,
456,
113,
457
],
[
223,
360,
225,
361
],
[
196,
360,
197,
361
],
[
240,
312,
241,
313
],
[
359,
287,
360,
288
],
[
240,
287,
241,
288
],
[
359,
260,
360,
262
],
[
358,
243,
360,
245
],
[
351,
233,
352,
234
],
[
336,
232,
338,
234
],
[
96,
232,
360,
784
]
] |
golffield
|
||
<image>Please carefully check the image and answer: The windmill on the upper right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The windmill on the upper right
|
[
[
640,
276,
644,
278
],
[
629,
232,
632,
235
],
[
579,
215,
580,
216
],
[
576,
200,
648,
275
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The oval ground track field. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The oval ground track field
|
[
[
631,
607,
632,
608
],
[
376,
575,
377,
576
],
[
319,
567,
320,
568
],
[
312,
567,
313,
568
],
[
267,
559,
268,
560
],
[
735,
541,
736,
542
],
[
751,
471,
752,
472
],
[
751,
468,
752,
470
],
[
208,
401,
209,
402
],
[
727,
400,
728,
401
],
[
711,
387,
712,
388
],
[
656,
360,
658,
362
],
[
647,
360,
648,
361
],
[
606,
352,
607,
353
],
[
595,
352,
601,
353
],
[
272,
328,
273,
330
],
[
352,
320,
354,
322
],
[
208,
320,
752,
616
]
] |
groundtrackfield
|
||
<image>Please carefully check the image and answer: A dark blue stadium in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A dark blue stadium in the middle
|
[] |
stadium
|
||
<image>Please carefully check the image and answer: The vehicle is below the vehicle on the top. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The vehicle is below the vehicle on the top
|
[
[
376,
463,
377,
464
],
[
407,
451,
408,
452
],
[
376,
400,
408,
464
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A vehicle is below the baseball field on the upper right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle is below the baseball field on the upper right
|
[
[
479,
725,
480,
726
],
[
448,
712,
480,
736
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A vehicle on the upper right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle on the upper right
|
[
[
647,
194,
648,
195
],
[
647,
191,
648,
192
],
[
600,
160,
648,
200
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: The windmill on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The windmill on the right
|
[
[
754,
542,
755,
544
],
[
775,
536,
776,
540
],
[
771,
536,
773,
544
],
[
724,
527,
725,
528
],
[
735,
512,
736,
513
],
[
688,
496,
770,
560
]
] |
windmill
|
||
<image>Please carefully check the image and answer: The white and gray vehicle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The white and gray vehicle
|
[
[
756,
775,
757,
776
],
[
720,
760,
760,
784
],
[
720,
760,
721,
761
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A basketball court in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A basketball court in the middle
|
[
[
420,
567,
421,
568
],
[
400,
566,
406,
568
],
[
351,
559,
352,
560
],
[
524,
527,
525,
528
],
[
534,
459,
536,
461
],
[
535,
449,
536,
450
],
[
352,
446,
353,
450
],
[
535,
439,
536,
440
],
[
368,
349,
369,
350
],
[
376,
287,
377,
288
],
[
540,
280,
541,
281
],
[
385,
256,
386,
257
],
[
344,
256,
560,
576
]
] |
basketballcourt
|
||
<image>Please carefully check the image and answer: The tennis court is on the lower left of the ground track field on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The tennis court is on the lower left of the ground track field on the right
|
[
[
399,
352,
400,
353
],
[
352,
352,
400,
384
]
] |
tenniscourt
|
||
<image>Please carefully check the image and answer: A orange baseball field. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A orange baseball field
|
[
[
424,
607,
425,
608
],
[
384,
560,
385,
561
],
[
384,
557,
385,
559
],
[
487,
531,
488,
532
],
[
478,
512,
480,
516
],
[
384,
496,
488,
608
]
] |
baseballfield
|
||
<image>Please carefully check the image and answer: The large overpass on the left. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The large overpass on the left
|
[
[
442,
448,
443,
449
],
[
439,
424,
440,
425
],
[
431,
400,
432,
401
],
[
183,
328,
184,
329
],
[
64,
328,
456,
504
]
] |
overpass
|
||
<image>Please carefully check the image and answer: The gray vehicle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
The gray vehicle
|
[
[
493,
480,
495,
483
],
[
480,
480,
496,
528
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A windmill on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A windmill on the right
|
[
[
683,
431,
686,
432
],
[
683,
408,
685,
409
],
[
656,
392,
704,
432
],
[
656,
392,
657,
393
]
] |
windmill
|
||
<image>Please carefully check the image and answer: A slender small bridge in the middle. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A slender small bridge in the middle
|
[
[
435,
583,
436,
584
],
[
310,
574,
312,
576
],
[
647,
519,
648,
520
],
[
635,
504,
637,
505
],
[
603,
504,
607,
506
],
[
495,
500,
496,
501
],
[
296,
488,
297,
489
],
[
192,
488,
647,
592
]
] |
bridge
|
||
<image>Please carefully check the image and answer: A vehicle on the upper right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A vehicle on the upper right
|
[
[
696,
215,
697,
216
],
[
750,
211,
752,
213
],
[
696,
200,
752,
232
]
] |
vehicle
|
||
<image>Please carefully check the image and answer: A bridge on the lower right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A bridge on the lower right
|
[
[
763,
599,
766,
600
],
[
760,
585,
800,
586
],
[
752,
584,
800,
600
]
] |
bridge
|
||
<image>Please carefully check the image and answer: A small bridge. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A small bridge
|
[
[
335,
482,
336,
483
],
[
351,
479,
352,
480
],
[
409,
463,
410,
464
],
[
415,
459,
416,
460
],
[
463,
441,
464,
442
],
[
503,
431,
504,
432
],
[
557,
415,
559,
416
],
[
371,
384,
372,
386
],
[
393,
376,
395,
377
],
[
475,
352,
477,
354
],
[
488,
348,
489,
349
],
[
575,
337,
576,
338
],
[
280,
328,
592,
488
]
] |
bridge
|
||
<image>Please carefully check the image and answer: A gray basketball court on the right. Based on your answer, detect all relevant objects in the image. Output each detected target's bbox coordinates in JSON format. The format of the bbox coordinates is:
```json
[
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]},
{"bbox_2d": [x1, y1, x2, y2], "keypoint1": [x3, y3], "keypoint2": [x4, y4]}
]
```
|
A gray basketball court on the right
|
[
[
644,
391,
645,
392
],
[
592,
379,
593,
380
],
[
735,
344,
736,
346
],
[
631,
304,
632,
305
],
[
584,
304,
736,
392
]
] |
basketballcourt
|
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 30
Size of downloaded dataset files:
1.11 GB
Size of the auto-converted Parquet files:
1.11 GB
Number of rows:
3,481