ลวิส

  • คำอธิบาย :

LVIS: ชุดข้อมูลสำหรับการแบ่งส่วนอินสแตนซ์คำศัพท์ขนาดใหญ่

  • เอกสารเพิ่มเติม : สำรวจเอกสารด้วยรหัส

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

  • โค้ดต้นฉบับ : tfds.datasets.lvis.Builder

  • เวอร์ชัน :

    • 1.1.0 : เพิ่มฟิลด์ neg_category_ids และ not_exhaustive_category_ids
    • 1.2.0 : เพิ่มชื่อคลาส
    • 1.3.0 : เพิ่มการแยกมินิวัล
    • 1.4.0 (ค่าเริ่มต้น) : เพิ่มมาสก์การแบ่งส่วนให้กับการแบ่งแบบมินิวัล
  • ขนาดดาวน์โหลด : 25.38 GiB

  • ขนาดชุดข้อมูล : 23.06 GiB

  • แคชอัตโนมัติ ( เอกสารประกอบ ): ไม่

  • สปลิต :

แยก ตัวอย่าง
'minival' 4,809
'test' 19,822
'train' 100,170
'validation' 19,809
  • โครงสร้างคุณสมบัติ :
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),
    }),
})
  • เอกสารประกอบคุณลักษณะ :
คุณสมบัติ ระดับ รูปร่าง ประเภท D คำอธิบาย
ฟีเจอร์ดิกต์
ภาพ ภาพ (ไม่มี, ไม่มี, 3) uint8
รูปภาพ/รหัส เทนเซอร์ int64
รหัสหมวดหมู่เชิงลบ ลำดับ (ClassLabel) (ไม่มี,) int64
รหัสหมวดหมู่ที่ไม่ครบถ้วน ลำดับ (ClassLabel) (ไม่มี,) int64
วัตถุ ลำดับ
วัตถุ/พื้นที่ เทนเซอร์ int64
วัตถุ/กล่อง ฟีเจอร์ BBox (4,) ลอย32
วัตถุ/รหัส เทนเซอร์ int64
วัตถุ/ฉลาก คลาสเลเบล int64
วัตถุ/การแบ่งส่วน ภาพ (ไม่มี, ไม่มี, 1) uint8

การสร้างภาพ

  • การอ้างอิง :
@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}
}