cmardb

  • 説明

このデータセットには、 - 手書きのバングラ数字 - 合計 6000 個のバングラ数字 (32x32 RGB カラー、6000 個の画像) のバランスのとれたデータセットの画像が含まれており、それぞれがクラス (桁ごと) ごとに 600 個の画像を持っています。手書きのデーヴァナーガリー数字 - 合計 3000 個のデーヴァナーガリー数字 (32x32 RGB カラー、3000 個の画像) のバランスのとれたデータセット。各クラス (桁ごと) に 300 個の画像があります。手書きのテルグ語数字 - 合計 3000 個のテルグ語数字 (32x32 RGB カラー、3000 個の画像) のバランスのとれたデータセット。各クラス (桁ごと) に 300 個の画像があります。

CMATERdb は、インドのジャダブプール大学の「トレーニング教育研究のためのマイクロプロセッサ アプリケーション センター」(CMATER) 研究室で作成されたパターン認識データベース リポジトリです。

FeaturesDict({
   
'image': Image(shape=(32, 32, 3), dtype=uint8),
   
'label': ClassLabel(shape=(), dtype=int64, num_classes=10),
})
  • 機能ドキュメント:
特徴クラスDタイプ説明
特徴辞書
画像画像(32、32、3) uint8
ラベルクラスラベルint64
@article{Das:2012:GAB:2161007.2161320,
  author
= {Das, Nibaran and Sarkar, Ram and Basu, Subhadip and Kundu, Mahantapas
           
and Nasipuri, Mita and Basu, Dipak Kumar},
  title
= {A Genetic Algorithm Based Region Sampling for Selection of Local Features
         
in Handwritten Digit Recognition Application},
  journal
= {Appl. Soft Comput.},
  issue_date
= {May, 2012},
  volume
= {12},
  number
= {5},
  month
= may,
  year
= {2012},
  issn
= {1568-4946},
  pages
= {1592--1606},
  numpages
= {15},
  url
= {http://dx.doi.org/10.1016/j.asoc.2011.11.030},
  doi
= {10.1016/j.asoc.2011.11.030},
  acmid
= {2161320},
  publisher
= {Elsevier Science Publishers B. V.},
  address
= {Amsterdam, The Netherlands, The Netherlands},
  keywords
= {Feature selection, Genetic algorithm, N-Quality consensus,
 
Optimal local regions, Region sampling, Variable sized local regions},
}
@article{Das:2012:SFC:2240301.2240421,
  author
= {Das, Nibaran and Reddy, Jagan Mohan and Sarkar, Ram and Basu, Subhadip and Kundu,
           
Mahantapas and Nasipuri, Mita and Basu, Dipak Kumar},
  title
= {A Statistical-topological Feature Combination for Recognition of Handwritten Numerals},
  journal
= {Appl. Soft Comput.},
  issue_date
= {August, 2012},
  volume
= {12},
  number
= {8},
  month
= aug,
  year
= {2012},
  issn
= {1568-4946},
  pages
= {2486--2495},
  numpages
= {10},
  url
= {http://dx.doi.org/10.1016/j.asoc.2012.03.039},
  doi
= {10.1016/j.asoc.2012.03.039},
  acmid
= {2240421},
  publisher
= {Elsevier Science Publishers B. V.},
  address
= {Amsterdam, The Netherlands, The Netherlands},
  keywords
= {Character recognition, Feature combination, MPCA, PCA, SVM, Statistical, Topological},
}

cmaterdb/bangla (デフォルト設定)

  • 構成の説明: CMATERdb バングラ数字

  • ダウンロードサイズ: 573.81 KiB

  • データセットのサイズ: 1.71 MiB

  • 分割:

スプリット
'test' 1,000
'train' 5,000

視覚化

cmardb/デヴァナガリ

  • 構成の説明: CMATERdb デヴァンガリ数字

  • ダウンロードサイズ: 275.29 KiB

  • データセットのサイズ: 869.50 KiB

  • 分割:

スプリット
'test' 500
'train' 2,500

視覚化

cmardb/テルグ語

  • 構成の説明: CMATERdb テルグ語数字

  • ダウンロードサイズ: 283.90 KiB

  • データセットのサイズ: 860.47 KiB

  • 分割:

スプリット
'test' 500
'train' 2,500

視覚化