Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ

Đếm đám đông với bài kiểm tra Bản đồ mật độ

Kiểm tra việc đếm bản đồ mật độ, tích hợp bản đồ, chú thích và hiệu ứng phối cảnh, số liệu đếm và giới hạn trong việc diễn giải ước tính đám đông.

Đường dẫn hướng dẫn liên quanBản đồ mật độ đếm đám đông
Câu hỏi 1 của 8

How does a density-map method commonly estimate the count in an image?

Xem trước mọi câu hỏi8 câu hỏi

Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.

  1. How does a density-map method commonly estimate the count in an image?
    • Sum or integrate the predicted spatial density values
    • Count the number of face identities stored in a database
    • Choose the largest local density value as the total
    • Multiply the image width by the number of detected cameras
  2. Why can density maps help in a tightly packed scene?
    • They remove all occlusion from the input image
    • They guarantee one exact coordinate for every person
    • They identify people even when no camera image exists
    • They can estimate spatial density without requiring a separate box for each visible person
  3. What does the sum of a predicted density map represent?
    • The total camera field of view in square meters
    • An estimated count under the model’s map and scaling conventions
    • An exact census of everyone in the venue
    • The number of annotated training images
  4. Why can perspective affect a density-map target?
    • Farther people always produce brighter pixels
    • A normalized kernel makes image geometry irrelevant
    • A person occupies different pixel areas depending on distance and camera geometry
    • Perspective changes the actual number of annotations
  5. What can a low count MAE fail to reveal?
    • Density may be misplaced in the image even when the total is close
    • Whether a model has any output values
    • Whether a scene contains image pixels
    • Whether a count can be represented as an integer
  6. How should training and test splits be designed for fixed camera footage?
    • Use the same annotated images for training and evaluation
    • Select test clips only after viewing model predictions
    • Separate scenes or cameras where appropriate to limit near-duplicate frame leakage
    • Randomly split adjacent video frames regardless of overlap
  7. Which claim can a density estimate support?
    • The identities and intentions of every person
    • A complete count beyond the camera’s field of view
    • Proof of an individual’s behavior
    • An approximate number of people visible in the camera’s covered scene
  8. What should be reported to make a crowd count interpretable?
    • The venue address without a count interval
    • The measurement definition, such as people per frame, zone, or time window
    • Only the model architecture name
    • The number of output-map pixels without units