Facial Recognition Quiz
Questions: 16 · 10 minutes
1. Why can people often recognize a familiar person even after that person changes their hairstyle?
Hair is normally the single most reliable identity feature.
A hairstyle change has no effect on how a face appears.
Recognition depends mainly on remembering the setting where the person was last seen.
Familiar-face recognition can draw on multiple features and their relationships, not just one surface detail.
2. In a facial recognition system, what is a face template?
A numerical representation derived from facial image features for comparison
A full-resolution photograph that must always be publicly displayed
A standard lighting setup used for every photograph
A list of names manually attached to a camera feed
3. Researchers train a recognition model and test it on different photos, but some photos of the same individuals appear in both sets. What is the main concern?
Identity leakage may make performance appear more generalizable than it really is.
The model can no longer produce similarity scores.
The test becomes too difficult because the identities overlap.
Every test result becomes a false non-match.
4. An eyewitness feels extremely confident when identifying a face seen briefly during a stressful event. What is the best conclusion?
A brief viewing is as informative as repeated familiarity.
High confidence guarantees accurate facial memory.
Stress always improves recognition of unfamiliar faces.
Confidence alone does not establish that the identification is accurate.
5. Software examines a face and labels the expression as a smile. Which task is this performing?
Facial-expression analysis rather than facial identity recognition
One-to-one identity verification
One-to-many identity identification
Face-template enrollment
6. At the same matching threshold, a face is searched against a much larger database. Why can false candidate matches become a greater concern?
Larger databases automatically lower the image resolution.
Every added identity changes the searched person’s facial features.
A larger database turns all comparisons into exact pixel matching.
More comparisons create more opportunities for an unrelated face to exceed the threshold.
7. A person enrolls in a facial recognition system under bright frontal lighting, then attempts to use it in dim side lighting. Which outcome becomes more plausible?
A false match caused solely by the person’s identity changing
A false non-match because the new image differs from the enrolled reference
Guaranteed recognition because lighting does not alter facial images
Expression detection instead of identity comparison
8. A facial recognition system will be used with people from several demographic groups. Which evaluation approach gives the clearest picture of whether performance is consistent?
Use one overall accuracy figure without subgroup results.
Test only the largest group because it has the most samples.
Report performance for each relevant group as well as overall performance.
Inspect a few successful examples from every group.
9. A program perfectly matches a photograph with an exact digital copy of that photograph. Why does this provide weak evidence that it can recognize the person in new images?
Identity can only be assessed from moving video.
A useful recognition system must assign a person’s name to every image.
The program may be matching identical pixels rather than identity across changes in pose, lighting, or expression.
Digital copies usually contain less facial information than printed photographs.
10. A system operator raises the similarity threshold required to declare a match. What tradeoff would generally be expected?
Both false matches and false non-matches must decrease.
False matches increase while false non-matches decrease.
The threshold changes processing speed but not error patterns.
False matches tend to decrease while false non-matches tend to increase.
11. What does prosopagnosia primarily refer to?
A general inability to see fine visual detail
A marked difficulty recognizing familiar faces that is not explained simply by poor eyesight
A tendency to forget people’s names while recognizing their faces
A temporary camera failure caused by low light
12. Why are upside-down faces often harder for people to recognize than upright faces?
Turning an image upside down removes facial features from it.
The visual system interprets every inverted face as an object without features.
Inversion disrupts the usual processing of facial configurations and feature relationships.
Only familiar faces can be viewed accurately upside down.
13. A security system asks a user to turn their head or blink before accepting a facial match. What is the main purpose of this step?
To identify the user’s emotional state
To improve the camera’s color balance
To help distinguish a live person from a presented photo or replayed recording
To convert a verification task into a database search
14. A mask covers a person’s nose and mouth. What is the most direct challenge this creates for face recognition?
Part of the facial information available for comparison is occluded.
The system automatically switches from verification to identification.
The mask changes the person’s identity.
The eyes become impossible to capture in an image.
15. A phone compares its current camera image only with the enrolled face of its owner. Which task is it performing?
Face detection without comparison
One-to-one identity verification
One-to-many identification
Facial-expression classification
16. What is the main distinction between face detection and face recognition?
Detection works on photographs, while recognition works only on video.
Detection locates the presence of a face, while recognition compares or associates it with an identity.
Detection analyzes expressions, while recognition measures image quality.
Detection requires a database, while recognition never uses stored references.