Wake_vision

  • Sự miêu tả :

Wake Vision là một tập dữ liệu lớn, chất lượng cao với hơn 6 triệu hình ảnh, vượt xa đáng kể quy mô và tính đa dạng của các tập dữ liệu tinyML hiện tại (100x). Tập dữ liệu này bao gồm các hình ảnh có chú thích cho biết mỗi hình ảnh có chứa một người hay không. Ngoài ra, nó còn kết hợp một tiêu chuẩn chi tiết toàn diện để đánh giá tính công bằng và chắc chắn, bao gồm giới tính nhận thức, độ tuổi nhận thức, khoảng cách chủ thể, điều kiện ánh sáng và mô tả. Nhãn Wake Vision có nguồn gốc từ các chú thích của Open Image được Google LLC cấp phép theo giấy phép CC BY 4.0. Các hình ảnh được liệt kê là có giấy phép CC BY 2.0. Lưu ý từ Hình ảnh mở: "mặc dù chúng tôi cố gắng xác định hình ảnh được cấp phép theo giấy phép Creative Commons Ghi công, chúng tôi không đưa ra tuyên bố hay bảo đảm nào về trạng thái giấy phép của từng hình ảnh và bạn nên tự mình xác minh giấy phép cho từng hình ảnh."

Tách ra Ví dụ
  • Cấu trúc tính năng :
FeaturesDict({
    'age_unknown': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'body_part': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'bright': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'dark': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'depiction': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'far': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'filename': Text(shape=(), dtype=string),
    'gender_unknown': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'medium_distance': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'middle_age': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'near': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'non-person_depiction': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'non-person_non-depiction': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'normal_lighting': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'older': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'person': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'person_depiction': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'predominantly_female': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'predominantly_male': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'young': ClassLabel(shape=(), dtype=int64, num_classes=2),
})
  • Tài liệu tính năng :
Tính năng Lớp học Hình dạng loại D Sự miêu tả
Tính năngDict
tuổi_không xác định Nhãn lớp int64
phần cơ thể Nhãn lớp int64
sáng Nhãn lớp int64
tối tăm Nhãn lớp int64
miêu tả Nhãn lớp int64
xa Nhãn lớp int64
tên tập tin Chữ sợi dây
giới tính_không xác định Nhãn lớp int64
hình ảnh Hình ảnh (Không, Không, 3) uint8
khoảng cách trung bình Nhãn lớp int64
tuổi trung niên Nhãn lớp int64
gần Nhãn lớp int64
không phải người_mô tả Nhãn lớp int64
không phải người_không miêu tả Nhãn lớp int64
normal_lighting Nhãn lớp int64
lớn hơn Nhãn lớp int64
người Nhãn lớp int64
người_mô tả Nhãn lớp int64
chủ yếu_nữ Nhãn lớp int64
chủ yếu_nam Nhãn lớp int64
trẻ Nhãn lớp int64
@article{banbury2024wake,
  title={Wake Vision: A Large-scale, Diverse Dataset and Benchmark Suite for TinyML Person Detection},
  author={Banbury, Colby and Njor, Emil and Stewart, Matthew and Warden, Pete and Kudlur, Manjunath and Jeffries, Nat and Fafoutis, Xenofon and Reddi, Vijay Janapa},
  journal={arXiv preprint arXiv:2405.00892},
  year={2024}
}