칼텍101

  • 설명 :

Caltech-101은 101개 클래스에 속하는 객체의 사진과 하나의 background clutter 클래스로 구성됩니다. 각 이미지에는 단일 개체로 라벨이 지정됩니다. 각 클래스에는 대략 40~800개의 이미지가 포함되어 있으며 총 9,000개의 이미지가 포함됩니다. 이미지는 크기가 다양하며 일반적인 가장자리 길이는 200~300픽셀입니다. 이 버전에는 이미지 수준 라벨만 포함되어 있습니다. 원본 데이터세트에는 경계 상자도 포함되어 있습니다.

나뉘다
'test' 6,084
'train' 3,060
  • 기능 구조 :
FeaturesDict({
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'image/file_name': Text(shape=(), dtype=string),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=102),
})
  • 기능 문서 :
특징 수업 모양 Dtype 설명
특징Dict
영상 영상 (없음, 없음, 3) uint8
이미지/파일_이름 텍스트
상표 클래스 라벨 정수64

심상

  • 인용 :
@article{FeiFei2004LearningGV,
  title={Learning Generative Visual Models from Few Training Examples: An Incremental Bayesian Approach Tested on 101 Object Categories},
  author={Li Fei-Fei and Rob Fergus and Pietro Perona},
  journal={Computer Vision and Pattern Recognition Workshop},
  year={2004},
}