en
0.25
0.5
0.75
1.25
1.5
1.75
2
Dense Active Appearance Models Using a Bounded Diameter Minimum Spanning Tree
Published on Oct 09, 20124126 Views
We present a method for producing dense Active Appearance Models (AAMs), suitable for video-realistic synthesis. To this end we estimate a joint alignment of all training images using a set of pairw
Related categories
Chapter list
Dense Active Appearance Models Using a Bounded Diameter Minimum Spanning Tree00:00
Active Appearance Models (AAMs)00:12
Dense AAMs (1)00:54
Dense AAMs (2)01:53
Dense AAMs for synthesis02:05
Prior work on automatic AAM building (1)02:50
Prior work on automatic AAM building (2)03:17
Approach for automatic AAM building03:21
Alignment through intermediate images (1)04:19
Alignment through intermediate images (2)05:23
Types of graph used for image registration05:41
Building a graph06:04
Registration through a graph (1)06:17
Registration through a graph (2)06:41
A minimum spanning tree over a training set07:28
Registration through a tree (1)07:53
Registration through a tree (2)08:36
Registration through a tree (3)09:04
Registration through a tree (4)09:08
Registration through a tree (5)09:18
Registration through a tree (6)09:23
The bounded diameter minimum spanning tree09:36
Effect of tree diameter on registration10:14
A bounded diameter minimum spanning tree11:22
Densification of an AAM given a joint alignment over the training images11:41
Approximation of flow fields11:56
Iterative densification (1)12:43
Iterative densification (2)13:22
Results of different densification approaches14:00
Summary14:50
All that’s left to say….15:14