কোকো

COCO হল একটি বড় মাপের অবজেক্ট ডিটেকশন, সেগমেন্টেশন এবং ক্যাপশনিং ডেটাসেট।

@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 (ডিফল্ট কনফিগারেশন)

  • কনফিগারেশনের বিবরণ : এই সংস্করণে 2014 সংস্করণের জন্য ছবি, বাউন্ডিং বক্স এবং লেবেল রয়েছে।

  • ডাউনলোড আকার : 37.57 GiB

  • ডেটাসেটের আকার : 37.35 GiB

  • বিভাজন :

বিভক্ত উদাহরণ
'test' 40,775
'test2015' ৮১,৪৩৪
'train' ৮২,৭৮৩
'validation' 40,504
  • বৈশিষ্ট্য গঠন :
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),
    }),
})
  • বৈশিষ্ট্য ডকুমেন্টেশন :
বৈশিষ্ট্য ক্লাস আকৃতি ডিটাইপ বর্ণনা
ফিচারসডিক্ট
ইমেজ ছবি (কোনটিই নয়, 3) uint8
ছবি/ফাইলের নাম পাঠ্য স্ট্রিং
ছবি/আইডি টেনসর int64
বস্তু ক্রম
বস্তু/ক্ষেত্র টেনসর int64
বস্তু/bbox বিবক্স ফিচার (৪,) float32
বস্তু/আইডি টেনসর int64
অবজেক্ট/ইজ_ভিড় টেনসর bool
বস্তু/লেবেল ক্লাসলেবেল int64

ভিজ্যুয়ালাইজেশন

coco/2017

  • কনফিগারের বিবরণ : এই সংস্করণে 2017 সংস্করণের জন্য ছবি, বাউন্ডিং বক্স এবং লেবেল রয়েছে।

  • ডাউনলোড আকার : 25.20 GiB

  • ডেটাসেটের আকার : 24.98 GiB

  • বিভাজন :

বিভক্ত উদাহরণ
'test' 40,670
'train' 118,287
'validation' 5,000
  • বৈশিষ্ট্য গঠন :
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),
    }),
})
  • বৈশিষ্ট্য ডকুমেন্টেশন :
বৈশিষ্ট্য ক্লাস আকৃতি ডিটাইপ বর্ণনা
ফিচারসডিক্ট
ইমেজ ছবি (কোনটিই নয়, 3) uint8
ছবি/ফাইলের নাম পাঠ্য স্ট্রিং
ছবি/আইডি টেনসর int64
বস্তু ক্রম
বস্তু/ক্ষেত্র টেনসর int64
বস্তু/bbox বিবক্স ফিচার (৪,) float32
বস্তু/আইডি টেনসর int64
অবজেক্ট/ইজ_ভিড় টেনসর bool
বস্তু/লেবেল ক্লাসলেবেল int64

ভিজ্যুয়ালাইজেশন

coco/2017_panoptic

  • কনফিগারের বিবরণ : এই সংস্করণে 2017 সংস্করণের জন্য ছবি, বাউন্ডিং বক্স এবং লেবেল রয়েছে।

  • ডাউনলোড আকার : 19.57 GiB

  • ডেটাসেটের আকার : 19.63 GiB

  • বিভাজন :

বিভক্ত উদাহরণ
'train' 118,287
'validation' 5,000
  • বৈশিষ্ট্য গঠন :
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),
    }),
})
  • বৈশিষ্ট্য ডকুমেন্টেশন :
বৈশিষ্ট্য ক্লাস আকৃতি ডিটাইপ বর্ণনা
ফিচারসডিক্ট
ইমেজ ছবি (কোনটিই নয়, 3) uint8
ছবি/ফাইলের নাম পাঠ্য স্ট্রিং
ছবি/আইডি টেনসর int64
প্যানোপটিক_ইমেজ ছবি (কোনটিই নয়, 3) uint8
panoptic_image/filename পাঠ্য স্ট্রিং
প্যানোপটিক_বস্তু ক্রম
প্যানোপটিক_বস্তু/ক্ষেত্র টেনসর int64
panoptic_objects/bbox বিবক্স ফিচার (৪,) float32
panoptic_objects/id টেনসর int64
panoptic_objects/is_crowd টেনসর bool
প্যানোপটিক_বস্তু/লেবেল ক্লাসলেবেল int64