引導式測驗 · 中等 等級

公共閉路電視測驗的人工智慧視訊分析

測試公共閉路電視分析:物件搜尋與身分、事件警報、本地驗證、隱私影響、存取日誌、保留和監督。

問題 1 的 8

影片查詢發現一個穿著紅色夾克的人。這個結果說明了什麼?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. 影片查詢發現一個穿著紅色夾克的人。這個結果說明了什麼?
    • The person committed the reported crime
    • The system recognized a unique face
    • A clip contains visual features matching the query, not a confirmed identity
    • The person appears in every connected camera
  2. 為什麼追蹤 ID 在擁擠的場景中不可靠?
    • The system may split one person into several tracks or merge different people
    • Each ID is a legally verified identity
    • Crowds prevent any video from being recorded
    • Tracking removes occlusion
  3. 在強光和下雨期間會出現即時密度警報。運營商應該做什麼?
    • Assume the alert is true because it is real time
    • Search every nearby person’s identity
    • Delete the original feed
    • Check the source feed and camera conditions before acting on the alert
  4. GAO 的公共技術評估對效益評估有何看法?
    • A demonstration proves crime reduction
    • Attribution can be difficult when multiple technologies and staff contribute
    • Video analytics always reduces crime
    • Public agencies do not need evaluation
  5. 在本地攝影機上部署事件偵測器之前,哪個測試最相關?
    • Use only the vendor’s clearest daytime examples
    • Measure model size and response time only
    • Evaluate representative footage across lighting, weather, camera types, and crowd conditions
    • Assume all camera feeds have identical quality
  6. Why check every frame when reviewing automated redaction?
    • The tool cannot process video
    • One blurred frame proves the whole clip is protected
    • Frame review establishes identity
    • A face or plate may be visible only briefly and missed by the detector
  7. What record helps reconstruct a search for public CCTV footage?
    • Camera ID, time range, query, model version, user, and source clip
    • Only the final summary
    • A vendor invoice
    • A list of every employee in the city
  8. Object detection and identity matching differ in which way?
    • They are the same operation under different names
    • Object detection locates a visual category; identity matching attempts to link a person to an identity
    • Object detection establishes probable cause
    • Identity matching cannot use images