0.25
0.5
0.75
1.25
1.5
1.75
2
Matching Point Sets with respect to the Earth Mover’s Distance
Published on Feb 25, 20073827 Views
Shape matching is a fundamental problem in computer vision: given two shapes A and B, one wants to determine how closely A resembles B, according to some distance measure between the shapes. In order
Chapter list
Matching Point Sets with respect to the Earth Mover’s Distance00:00
Today00:42
Earth Mover’s Distance01:16
Earth Mover’s Distance01:56
Earth Mover’s Distance03:45
Earth Mover’s Distance04:51
Earth Mover’s Distance05:42
EMD for Shape Matching07:18
EMD for Shape Matching07:32
EMD for Shape Matching07:55
EMD for Shape Matching07:59
EMD for Pattern Matching08:30
EMD for Pattern Matching09:28
EMD for Pattern Matching10:11
The Goal10:33
The Goal11:02
The Goal11:19
EMD: the good, the bad, and the ugly12:41
A simple case: Fermat-Weber14:53
Today16:12
Some simple deas about the results16:19
A couple of lemmas17:35
A couple of lemmas18:42
Refine each neighbourhood19:13
Refine each neighbourhood20:06
Many pairs: probabilistic improvement20:42
Same weight: simpler and quicker22:07
Summary22:57
Thanks23:44