Stereo Depth Estimation Quiz

Answers and explanations appear as you take the quiz.

1. What is 'disparity' in stereo vision?
   • The horizontal shift of a matching point between the left and right images
   • The brightness difference between the two cameras
   • The vertical distance between the two cameras
   • The color mismatch caused by lens differences
2. How does depth relate to disparity?
   • Depth is directly proportional to disparity
   • Depth is inversely proportional to disparity
   • Depth equals disparity times the baseline
   • Depth is unrelated to disparity
3. Why are images 'rectified' before matching?
   • To increase the image resolution
   • To remove color from the images
   • To make corresponding points lie on the same horizontal row
   • To blur out moving objects
4. Which scenario is hardest for stereo matching?
   • A richly textured outdoor scene
   • A close-up of a patterned rug
   • A scene with strong, varied edges
   • A blank white wall with no texture
5. What does the 'baseline' refer to?
   • The physical distance between the two cameras
   • The minimum detectable depth
   • The reference frame's exposure time
   • The lowest brightness in the image
6. What is a 'cost volume' in deep stereo networks?
   • A measure of GPU memory usage
   • A 3D structure scoring how well features match at each candidate disparity
   • The dollar cost of the camera hardware
   • A list of detected object categories
7. Why are depth estimates often noisier for faraway objects than nearby ones?
   • Far objects always occupy more pixels than nearby objects
   • Stereo depth is calculated from color saturation rather than disparity
   • Far objects have small disparity, so a small disparity error can produce a larger depth error
   • The camera baseline increases automatically with object distance
8. What can LiDAR add when it is fused with stereo depth estimation?
   • A replacement for all image data in every scene
   • A guarantee that stereo estimates have no errors
   • A way to remove the need for camera calibration
   • An additional depth measurement that can help constrain stereo matching