stanford_dogs

  • 説明:

Stanford Dogs データセットには、世界中の 120 種類の犬の画像が含まれています。このデータセットは、画像を細かく分類するタスクのために、ImageNet の画像と注釈を使用して構築されています。 20,580 枚の画像があり、そのうち 12,000 枚がトレーニングに使用され、8580 枚がテストに使用されます。クラス ラベルとバウンディング ボックスの注釈は、12,000 枚の画像すべてに提供されます。

スプリット
'test' 8,580
'train' 12,000
  • 機能構造:
FeaturesDict({
   
'image': Image(shape=(None, None, 3), dtype=uint8),
   
'image/filename': Text(shape=(), dtype=string),
   
'label': ClassLabel(shape=(), dtype=int64, num_classes=120),
   
'objects': Sequence({
       
'bbox': BBoxFeature(shape=(4,), dtype=float32),
   
}),
})
  • 機能のドキュメント:
特徴クラスDtype説明
特徴辞書
画像画像(なし、なし、3) uint8
画像/ファイル名文章ストリング
ラベルクラスラベルint64
オブジェクト順序
オブジェクト/bbox BBoxFeature (4) float32

視覚化

  • 引用
@inproceedings{KhoslaYaoJayadevaprakashFeiFei_FGVC2011,
author
= "Aditya Khosla and Nityananda Jayadevaprakash and Bangpeng Yao and
          Li Fei-Fei"
,
title
= "Novel Dataset for Fine-Grained Image Categorization",
booktitle
= "First Workshop on Fine-Grained Visual Categorization,
             IEEE Conference on Computer Vision and Pattern Recognition"
,
year
= "2011",
month
= "June",
address
= "Colorado Springs, CO",
}
@inproceedings{imagenet_cvpr09,
        AUTHOR
= {Deng, J. and Dong, W. and Socher, R. and Li, L.-J. and
                 
Li, K. and Fei-Fei, L.},
        TITLE
= { {ImageNet: A Large-Scale Hierarchical Image Database} },
        BOOKTITLE
= {CVPR09},
        YEAR
= {2009},
        BIBSOURCE
= "http://www.image-net.org/papers/imagenet_cvpr09.bib"}