デイビス

  • 説明:

DAVIS 2017 ビデオ オブジェクト セグメンテーション データセット。

スプリット
'train' 60
'validation' 30
  • 機能構造:
FeaturesDict({
    'metadata': FeaturesDict({
        'num_frames': int64,
        'video_name': string,
    }),
    'video': Sequence({
        'frames': Image(shape=(None, None, 3), dtype=uint8),
        'segmentations': Image(shape=(None, None, 1), dtype=uint8),
    }),
})
  • 機能のドキュメント:
特徴クラスDtype説明
特徴辞書
メタデータ特徴辞書
メタデータ/num_framesテンソルint64
メタデータ/ビデオ名テンソルストリング
ビデオ順序
ビデオ/フレーム画像(なし、なし、3) uint8
ビデオ/セグメンテーション画像(なし、なし、1) uint8
@article{DBLP:journals/corr/Pont-TusetPCASG17,
  author    = {Jordi Pont{-}Tuset and
               Federico Perazzi and
               Sergi Caelles and
               Pablo Arbelaez and
               Alexander Sorkine{-}Hornung and
               Luc Van Gool},
  title     = {The 2017 {DAVIS} Challenge on Video Object Segmentation},
  journal   = {CoRR},
  volume    = {abs/1704.00675},
  year      = {2017},
  url       = {http://arxiv.org/abs/1704.00675},
  archivePrefix = {arXiv},
  eprint    = {1704.00675},
  timestamp = {Mon, 13 Aug 2018 16:48:55 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/Pont-TusetPCASG17.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

davis/480p (デフォルト設定)

  • 構成の説明: データセットの 480p バージョン

  • ダウンロードサイズ: 794.19 MiB

  • データセットサイズ: 792.26 MiB

  • ( tfds.as_dataframe ):

デイビス/フル解像度

  • 構成の説明: データセットのフル解像度バージョン。

  • ダウンロードサイズ: 2.75 GiB

  • データセットサイズ: 2.78 GiB

  • ( tfds.as_dataframe ):