The minimum circle is more expensive to compute (for an irregular polygon) but that's O(n). Since the circle is smaller it will be no worse and maybe a good bit better for weeding out non-overlaps and the bonus that brings will depend entirely on the dataset being examined. Without seeing sample data, would you take the O(n) hit for an O(n2) gain? I probably would.
(I bet it's not)
It's a gamble either way. :-)
In reply to Re^3: searching polygons not merged
by hippo
in thread searching polygons not merged
by dideod.yang
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |