Quiz sur la reconnaissance faciale

Les réponses et les explications apparaissent au fur et à mesure que vous répondez au quiz.

1. A phone compares a live face with one enrolled image. Which task is this?
   • One-to-one verification
   • One-to-many gallery identification
   • Reconnaissance des émotions
   • Segmentation d'images
2. A search compares one probe image against a large collection. What task is being performed?
   • One-to-one verification
   • One-to-many identification
   • Face detection only
   • Age estimation
3. Why should gallery size be considered when evaluating a search result?
   • Gallery size fixes image quality
   • More candidates prove identity
   • A low false-match rate per comparison does not fully describe a large-gallery search
   • It makes privacy review unnecessary
4. Which conditions can affect facial-comparison performance?
   • Only the file extension
   • The subject’s profession
   • The time of day the system was installed
   • Lighting, pose, image resolution and age differences between images
5. Why report false matches and false non-matches separately?
   • They always occur at the same rate
   • They represent different errors with different consequences
   • One metric proves identity
   • They measure emotion
6. A system returns a candidate list. How should it be used in a consequential process?
   • As a conclusive identity declaration
   • As proof of intent
   • As a lead for appropriate corroboration and review
   • As a substitute for privacy controls
7. What does a similarity result establish?
   • A comparison measurement under specified conditions, not identity certainty or intent
   • A person’s character
   • A person’s emotional state
   • A verified legal identity in every setting
8. Which safeguards belong in a deployment review?
   • Only match speed
   • Only the camera brand
   • Publicly sharing the gallery
   • Collection, retention, consent, access and challenge procedures