Computer Vision Foundations · free preview

Assumptions, Accidents, and What Images Hide

The generic-view assumption and its everyday violations, support cues, raindrop light fields, and the many things an image does not show.

1. Smooth continuations lie

Smooth continuations lie

2. Cues for support

Cues for support

3. Raindrops as light field cameras

Raindrops as light field cameras

4. Plato's cave

Plato's cave

5. How do you know something is wet?

How do you know something is wet?

6. When continuity fails

A vision algorithm groups every smooth curve in a photograph into a single object, using the principle of continuity. On a busy portrait it merges a finger contour with the outline of a lip because the two curves align perfectly. What does this failure teach?

7. Which cue wins on support

A doctored photograph shows a ball with its drop shadow far to the side while a slack string drapes from the ball to the ground surface. The two support cues disagree. Which reading does the combination most strongly support, and why?

Practice this interactively

Free account · instant grading · spaced review that schedules itself.

Start this course for free