کوکو

COCO یک مجموعه داده شناسایی، تقسیم بندی و زیرنویس در مقیاس بزرگ است.

@article{DBLP:journals/corr/LinMBHPRDZ14,
  author    = {Tsung{-}Yi Lin and
               Michael Maire and
               Serge J. Belongie and
               Lubomir D. Bourdev and
               Ross B. Girshick and
               James Hays and
               Pietro Perona and
               Deva Ramanan and
               Piotr Doll{'{a} }r and
               C. Lawrence Zitnick},
  title     = {Microsoft {COCO:} Common Objects in Context},
  journal   = {CoRR},
  volume    = {abs/1405.0312},
  year      = {2014},
  url       = {http://arxiv.org/abs/1405.0312},
  archivePrefix = {arXiv},
  eprint    = {1405.0312},
  timestamp = {Mon, 13 Aug 2018 16:48:13 +0200},
  biburl    = {https://dblp.org/rec/bib/journals/corr/LinMBHPRDZ14},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

coco/2014 (پیکربندی پیش فرض)

  • توضیحات پیکربندی : این نسخه شامل تصاویر، کادرهای محدود کننده و برچسب‌هایی برای نسخه 2014 است.

  • حجم دانلود : 37.57 GiB

  • حجم مجموعه داده : 37.35 GiB

  • تقسیم ها :

شکاف مثال ها
'test' 40775
'test2015' 81,434
'train' 82783
'validation' 40,504
  • ساختار ویژگی :
FeaturesDict({
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'image/filename': Text(shape=(), dtype=string),
    'image/id': int64,
    'objects': Sequence({
        'area': int64,
        'bbox': BBoxFeature(shape=(4,), dtype=float32),
        'id': int64,
        'is_crowd': bool,
        'label': ClassLabel(shape=(), dtype=int64, num_classes=80),
    }),
})
  • مستندات ویژگی :
ویژگی کلاس شکل نوع D شرح
FeaturesDict
تصویر تصویر (هیچ، هیچ، 3) uint8
تصویر/نام فایل متن رشته
تصویر / شناسه تانسور int64
اشیاء توالی
اشیاء/منطقه تانسور int64
اشیاء/bbox ویژگی BBox (4،) float32
اشیاء / شناسه تانسور int64
اشیاء/is_crowd تانسور بوول
اشیاء/برچسب ClassLabel int64

تجسم

coco/2017

  • توضیحات پیکربندی : این نسخه شامل تصاویر، کادرهای مرزبندی و برچسب‌هایی برای نسخه 2017 است.

  • حجم دانلود : 25.20 GiB

  • حجم مجموعه داده : 24.98 GiB

  • تقسیم ها :

شکاف مثال ها
'test' 40,670
'train' 118287
'validation' 5000
  • ساختار ویژگی :
FeaturesDict({
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'image/filename': Text(shape=(), dtype=string),
    'image/id': int64,
    'objects': Sequence({
        'area': int64,
        'bbox': BBoxFeature(shape=(4,), dtype=float32),
        'id': int64,
        'is_crowd': bool,
        'label': ClassLabel(shape=(), dtype=int64, num_classes=80),
    }),
})
  • مستندات ویژگی :
ویژگی کلاس شکل نوع D شرح
FeaturesDict
تصویر تصویر (هیچ، هیچ، 3) uint8
تصویر/نام فایل متن رشته
تصویر / شناسه تانسور int64
اشیاء توالی
اشیاء/منطقه تانسور int64
اشیاء/bbox ویژگی BBox (4،) float32
اشیاء / شناسه تانسور int64
اشیاء/is_crowd تانسور بوول
اشیاء/برچسب ClassLabel int64

تجسم

coco/2017_panoptic

  • توضیحات پیکربندی : این نسخه شامل تصاویر، کادرهای مرزبندی و برچسب‌هایی برای نسخه 2017 است.

  • حجم دانلود : 19.57 GiB

  • حجم مجموعه داده : 19.63 GiB

  • تقسیم ها :

شکاف مثال ها
'train' 118287
'validation' 5000
  • ساختار ویژگی :
FeaturesDict({
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'image/filename': Text(shape=(), dtype=string),
    'image/id': int64,
    'panoptic_image': Image(shape=(None, None, 3), dtype=uint8),
    'panoptic_image/filename': Text(shape=(), dtype=string),
    'panoptic_objects': Sequence({
        'area': int64,
        'bbox': BBoxFeature(shape=(4,), dtype=float32),
        'id': int64,
        'is_crowd': bool,
        'label': ClassLabel(shape=(), dtype=int64, num_classes=133),
    }),
})
  • مستندات ویژگی :
ویژگی کلاس شکل نوع D شرح
FeaturesDict
تصویر تصویر (هیچ، هیچ، 3) uint8
تصویر/نام فایل متن رشته
تصویر / شناسه تانسور int64
panoptic_image تصویر (هیچ، هیچ، 3) uint8
panoptic_image/نام فایل متن رشته
panoptic_objects توالی
panoptic_objects/ area تانسور int64
panoptic_objects/bbox ویژگی BBox (4،) float32
panoptic_objects/id تانسور int64
panoptic_objects/is_crowd تانسور بوول
panoptic_objects/label ClassLabel int64