Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ
Bài kiểm tra tái nhận dạng người qua camera
Kiểm tra khả năng truy xuất qua nhiều camera, giới hạn xuất hiện, thăm dò so với thư viện, các trường hợp không khớp được đặt mở, số liệu xếp hạng, độ không chắc chắn về danh tính và rủi ro về quyền riêng tư.
Câu hỏi 1 của 8
Which task does person re-identification address in the guide?
Xem trước mọi câu hỏi8 câu hỏi
Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.
- Which task does person re-identification address in the guide?
- Following a person only through consecutive frames of one camera
- Reading a person's legal name from every video frame
- Retrieving possible sightings of the same person across different camera views
- Determining whether any movement is a crime
- A system returns several gallery images ranked for a probe. What does the top rank establish?
- It is the system's most similar candidate under that setup, not proof of identity
- The gallery definitely contains the queried person
- The person's legal identity has been verified
- All lower-ranked candidates are unrelated by definition
- Why might two different people in similar uniforms be falsely linked?
- A model always reads names from clothing
- A high similarity score proves both people share a route
- Non-overlapping cameras cannot record similar colors
- Appearance features can resemble each other despite different identities
- A query person's image is absent from the gallery. What capability does a real-world evaluation need?
- A rule that forces a match to the nearest candidate
- A way to reject or qualify a no-match result
- A guarantee that every gallery contains the target
- A larger similarity score reported as certainty
- What does rank-one accuracy count in a specified re-ID benchmark?
- Queries whose first-ranked candidate is a correct match
- Every relevant gallery image at any rank
- The probability that a candidate has a known legal name
- The share of public cameras that use the same model
- Why can a strong benchmark rank-one score fail to predict real deployment error?
- Mean average precision has the same value in all datasets
- Benchmark labels automatically transfer to every new camera
- Gallery composition, camera conditions and no-match cases may differ
- Rank-one scoring includes the privacy purpose of every deployment
- What privacy risk can arise even when the system never attaches a person's name?
- Only a spelling error in the gallery file name
- Linking sightings can reveal routes, routines and associations
- A guarantee that all video becomes public
- No privacy risk, because names were omitted
- A transit agency only needs station-level flow counts. What question should precede individual re-ID?
- Which vendor offers the largest possible gallery?
- How long can every person's route be retained?
- Can a high similarity score authorize broader use?
- Can less identifying aggregate counts meet the purpose?