리비스

  • 설명 :

LVIS: 대규모 어휘 인스턴스 분할을 위한 데이터 세트.

  • 추가 문서 : 코드 있는 논문에서 탐색

  • 홈페이지 : https://www.lvisdataset.org/

  • 소스 코드 : tfds.datasets.lvis.Builder

  • 버전 :

    • 1.1.0 : neg_category_idsnot_exhaustive_category_ids 필드가 추가되었습니다.
    • 1.2.0 : 클래스 이름이 추가되었습니다.
    • 1.3.0 : 미니벌 분할이 추가되었습니다.
    • 1.4.0 (기본값) : 미니벌 분할에 분할 마스크가 추가되었습니다.
  • 다운로드 크기 : 25.38 GiB

  • 데이터 세트 크기 : 23.06 GiB

  • 자동 캐시 ( 문서 ): 아니요

  • 분할 :

나뉘다 예시
'minival' 4,809
'test' 19,822
'train' 100,170
'validation' 19,809
  • 기능 구조 :
FeaturesDict({
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'image/id': int64,
    'neg_category_ids': Sequence(ClassLabel(shape=(), dtype=int64, num_classes=1203)),
    'not_exhaustive_category_ids': Sequence(ClassLabel(shape=(), dtype=int64, num_classes=1203)),
    'objects': Sequence({
        'area': int64,
        'bbox': BBoxFeature(shape=(4,), dtype=float32),
        'id': int64,
        'label': ClassLabel(shape=(), dtype=int64, num_classes=1203),
        'segmentation': Image(shape=(None, None, 1), dtype=uint8),
    }),
})
  • 기능 설명서 :
특징 수업 모양 디타입 설명
특징 사전
영상 영상 (없음, 없음, 3) uint8
이미지/아이디 텐서 int64
부정_카테고리_아이디 시퀀스(클래스 레이블) (없음,) int64
완전하지 않은 카테고리 ID 시퀀스(클래스 레이블) (없음,) int64
사물 순서
객체/영역 텐서 int64
객체/bbox BBox 기능 (4,) 플로트32
객체/아이디 텐서 int64
객체/레이블 클래스 라벨 int64
객체/세분화 영상 (없음, 없음, 1) uint8

심상

  • 인용 :
@inproceedings{gupta2019lvis,
  title={ {LVIS}: A Dataset for Large Vocabulary Instance Segmentation},
  author={Gupta, Agrim and Dollar, Piotr and Girshick, Ross},
  booktitle={Proceedings of the {IEEE} Conference on Computer Vision and Pattern Recognition},
  year={2019}
}