en
0.25
0.5
0.75
1.25
1.5
1.75
2
Bag Dissimilarities for Multiple Instance Learning
Published on Oct 17, 20113335 Views
When objects cannot be represented well by single feature vectors, a collection of feature vectors can be used. This is what is done in Multiple Instance learning, where it is called a bag of instance
Related categories
Chapter list
Bag Dissimilarities for Multiple Instance Learning00:00
How to label this image?00:52
... “Red chilli”? - 101:42
... “Red chilli”? - 201:45
How to REPRESENT this image?01:56
Multiple Instance Learning (MIL) - 102:05
Contents03:00
Multiple Instance Learning (MIL) - 203:33
Multiple Instance Learning (MIL) - 304:19
Multiple Instance Learning (MIL) - 404:59
Multiple Instance Learning (MIL) - 505:39
Multiple Instance Learning (MIL) - 606:02
Notation06:25
Bag dissimilarities07:04
Bag dissim. using pairwise dist.08:32
Bag distribution dissimilarities08:48
Does it make sense?10:18
Results - 111:51
Results - 213:09
Zoom in (1)14:45
Zoom in (2)15:39
Min-min distance? - 117:13
Min-min distance? - 217:27
Min-min distance? - 317:48
Mean-min. distance?18:35
Earth movers distance?18:55
More results19:15
Zoomed...19:23
Are there different MIL problems?19:37
Conclusions20:38