引導式測驗 · 中等 等級
偽造影像中的陰影與光線線索測驗
Test cast-shadow geometry, perspective and multiple-light alternatives, source images, physical assumptions, and AI-origin limits.
問題 1 的 8
What should be established before comparing two cast-shadow directions?
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。
- What should be established before comparing two cast-shadow directions?
- Whether objects share a comparable surface and lighting setup
- Only the image file size
- Only the number of shadows
- Only the caption wording
- Why can two real indoor shadows point differently?
- Real light never makes shadows
- Perspective proves editing
- Several light sources or bounced light may be present
- Every indoor image is AI-generated
- What can camera perspective do to parallel scene directions?
- Erase every shadow
- Make them appear to converge or diverge in the image
- Prove a fake scene
- Remove the need for geometry
- A shadow seems inconsistent under a single-sun model. What is a careful conclusion?
- The image was certainly AI-generated
- The photographer must be dishonest
- No other light or surface is possible
- The clue merits further analysis under stated assumptions
- Why inspect contact points and surfaces?
- They constrain how object and shadow geometry relate
- They reveal the camera serial number
- They determine citation style
- They guarantee one sun
- What else should be sought before a public fake-image claim?
- Only a brighter screenshot
- Only a model confidence label
- Original file history and independent event evidence
- Only one line drawn on a thumbnail
- What does a physically plausible shadow fail to establish?
- That light could exist
- That the image is genuine and unedited
- That a shadow can be measured
- That geometry can be tested
- Which cause can shadow analysis alone not uniquely identify?
- Whether a contact point exists
- Whether a shadow has an endpoint
- Whether a surface has slope
- AI generation versus compositing versus unmodeled real lighting