कोको

COCO एक बड़े पैमाने पर ऑब्जेक्ट डिटेक्शन, सेगमेंटेशन और कैप्शनिंग डेटासेट है।

@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}
}

कोको/2014 (डिफ़ॉल्ट कॉन्फ़िगरेशन)

  • कॉन्फ़िगरेशन विवरण : इस संस्करण में 2014 संस्करण के लिए चित्र, बाउंडिंग बॉक्स और लेबल शामिल हैं।

  • डाउनलोड का आकार : 37.57 GiB

  • डेटासेट का आकार : 37.35 GiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 40,775
'test2015' 81,434
'train' 82,783
'validation' 40,504
  • फ़ीचर संरचना :
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),
    }),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
छवि छवि (कोई नहीं, कोई नहीं, 3) uint8
छवि/फ़ाइल नाम मूलपाठ डोरी
छवि/आईडी टेन्सर int64
वस्तुओं क्रम
वस्तुओं / क्षेत्र टेन्सर int64
ऑब्जेक्ट्स/बीबॉक्स बीबॉक्स फीचर (4,) फ्लोट32
वस्तुओं/आईडी टेन्सर int64
ऑब्जेक्ट्स/is_crowd टेन्सर बूल
ऑब्जेक्ट्स / लेबल क्लासलेबल int64

VISUALIZATION

कोको/2017

  • कॉन्फ़िगरेशन विवरण : इस संस्करण में 2017 संस्करण के लिए चित्र, बाउंडिंग बॉक्स और लेबल शामिल हैं।

  • डाउनलोड आकार : 25.20 GiB

  • डेटासेट का आकार : 24.98 GiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 40,670
'train' 118,287
'validation' 5,000
  • फ़ीचर संरचना :
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),
    }),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
छवि छवि (कोई नहीं, कोई नहीं, 3) uint8
छवि/फ़ाइल नाम मूलपाठ डोरी
छवि/आईडी टेन्सर int64
वस्तुओं क्रम
वस्तुओं / क्षेत्र टेन्सर int64
ऑब्जेक्ट्स/बीबॉक्स बीबॉक्स फीचर (4,) फ्लोट32
वस्तुओं/आईडी टेन्सर int64
ऑब्जेक्ट्स/is_crowd टेन्सर बूल
ऑब्जेक्ट्स / लेबल क्लासलेबल int64

VISUALIZATION

कोको/2017_पैनोप्टिक

  • कॉन्फ़िगरेशन विवरण : इस संस्करण में 2017 संस्करण के लिए चित्र, बाउंडिंग बॉक्स और लेबल शामिल हैं।

  • डाउनलोड का आकार : 19.57 GiB

  • डेटासेट का आकार : 19.63 GiB

  • विभाजन :

विभाजित करना उदाहरण
'train' 118,287
'validation' 5,000
  • फ़ीचर संरचना :
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),
    }),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
छवि छवि (कोई नहीं, कोई नहीं, 3) uint8
छवि/फ़ाइल नाम मूलपाठ डोरी
छवि/आईडी टेन्सर int64
panoptic_image छवि (कोई नहीं, कोई नहीं, 3) uint8
panoptic_image/filename मूलपाठ डोरी
panoptic_objects क्रम
panoptic_objects/area टेन्सर int64
panoptic_objects/bbox बीबॉक्स फीचर (4,) फ्लोट32
panoptic_objects/id टेन्सर int64
panoptic_objects/is_crowd टेन्सर बूल
पैनोप्टिक_ऑब्जेक्ट्स/लेबल क्लासलेबल int64