당뇨병_망막병증_탐지

  • 설명 :

다양한 이미징 조건에서 촬영한 대규모 고해상도 망막 이미지 세트.

나뉘다
'sample' 10
'test' 42,670
'train' 35,126
'validation' 10,906
  • 기능 구조 :
FeaturesDict({
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=5),
    'name': Text(shape=(), dtype=string),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
영상 영상 (없음, 없음, 3) uint8
상표 클래스 레이블 int64
이름 텍스트
@ONLINE {kaggle-diabetic-retinopathy,
    author = "Kaggle and EyePacs",
    title  = "Kaggle Diabetic Retinopathy Detection",
    month  = "jul",
    year   = "2015",
    url    = "https://www.kaggle.com/c/diabetic-retinopathy-detection/data"
}

diabetic_retinopathy_detection/original(기본 구성)

  • 구성 설명 : 원래 해상도 및 품질의 이미지입니다.

  • 데이터세트 크기 : 89.15 GiB

  • 그림 ( tfds.show_examples ):

심상

diabetic_retinopathy_detection/1M

  • 구성 설명 : 이미지에는 72 품질의 약 1,000,000픽셀이 있습니다.

  • 데이터세트 크기 : 3.96 GiB

  • 그림 ( tfds.show_examples ):

심상

diabetic_retinopathy_detection/250K

  • 구성 설명 : 이미지는 약 250,000픽셀, 72 품질입니다.

  • 데이터세트 크기 : 1.30 GiB

  • 그림 ( tfds.show_examples ):

심상

diabetic_retinopathy_detection/btgraham-300

  • 구성 설명 : 이미지는 2015년 Kaggle 대회 우승자처럼 전처리되었습니다. 먼저 안구 반경이 300픽셀이 되도록 크기를 조정한 다음 반경의 90%로 자르고 마지막으로 다음과 같이 인코딩합니다. 72 JPEG 품질.

  • 데이터세트 크기 : 3.65 GiB

  • 그림 ( tfds.show_examples ):

심상