A sliding-window detector does not only need boxes of different sizes. It needs different shapes too: a standing person fills a tall, narrow box, while a car seen side-on fills a wide, flat one.
Run a sliding window on a single photo pixels wide and pixels tall, trying one box area in three shapes:
| box | width | height |
|---|---|---|
| square | ||
| wide | ||
| tall |
Every placement of every box is one run of the classifier.
How many classifier runs does this one photo cost?
Select all that apply.