Guide-linked quiz · Medium Level

AI in Colorizing Historical Photos and Film Quiz

Learn how colorization models infer missing hues, stabilize film frames, and distinguish artistic reconstruction from historical evidence.

Question 1 of 8

Why must colorization AI 'guess' the colors in an old photo?

Preview every question8 questions

Answers and explanations appear as you take the quiz.

  1. Why must colorization AI 'guess' the colors in an old photo?
    • Because black-and-white images record only brightness, not actual color
    • Because the camera deleted the colors on purpose
    • Because old film was always painted gray
    • Because color always fades to black over time
  2. In the Colorful Image Colorization paper, what does the model predict from the L channel?
    • Pure RGB with all channels redrawn
    • The Lab color space, predicting only the color channels
    • CMYK for printing
    • Hexadecimal only
  3. How are colorization models typically trained?
    • By hand-painting millions of images
    • By scanning paint catalogs
    • By converting color photos to grayscale and learning to restore the color
    • By copying colors from nearby pixels only
  4. What technique did DeOldify popularize for more believable colors?
    • Random color assignment
    • Manual frame-by-frame painting
    • Black-and-white sharpening only
    • A GAN-style generator and critic judging realism
  5. What extra challenge does colorizing film add compared to a single photo?
    • Maintaining temporal consistency so colors do not flicker between frames
    • Films have no brightness data
    • Films are always already in color
    • You can only colorize the first frame
  6. Why is AI colorization considered controversial for archival accuracy?
    • It makes files too small
    • The colors are plausible guesses, not verified historical truth
    • It always turns everything purple
    • It erases the original photo permanently
  7. What can historical metadata contribute to a colorization project?
    • A guarantee that every predicted pixel is historically exact
    • The original color values from a grayscale scan
    • Evidence that can constrain plausible colors, such as documented uniforms or paint schemes
    • A replacement for examining the source photograph
  8. How should an archival release distinguish AI colorization from verified historical color?
    • Present it as the original color photograph
    • Remove the grayscale source from the archive
    • Call every plausible hue a verified fact
    • Label it as a colorized reconstruction and document the evidence used

Keep testing yourself

More quizzes picked for your level.

View all quizzes