en
0.25
0.5
0.75
1.25
1.5
1.75
2
MoT - Mixture of Trees Probabilistic Graphical Model for Video Segmentation
Published on Oct 09, 20123326 Views
We present a novel mixture of trees (MoT) graphical model for video segmentation. Each component in this mixture represents a tree structured temporal linkage between super-pixels from the first to
Related categories
Chapter list
MoT – Mixture of Trees Probabilistic Graphical Model for Video Segmentation00:00
Video Segmentation (1)00:16
Video Segmentation (2)00:23
Our framework (1)00:46
Our framework (2)00:53
Road map (1)02:15
Road map (2)02:23
Related work (1)02:25
Related work (2)02:47
GeoS: Geodesic Image Segmentation02:54
Related work (3)03:26
Motion coherent tracking with multi-label MRF optimization03:36
Related work (4)04:18
Combining Self Training and Active Learning for Video Segmentation04:22
Related work (5)05:10
Related work (6)05:13
Road map (3)05:22
Model - Overview05:24
Model – Input/Output06:06
Model – Pair of frames06:31
Model – Appearance layer07:14
Model – Label layer07:57
Model – Mapping variable08:16
Model – Mixture of Trees (1)08:38
Model – Mixture of Trees (2)08:56
Model – Mixture of Trees (3)08:59
Model – Full model09:07
Road map (4)09:14
Inference – Mixture of Trees graphical model09:16
Inference – Structured variational inference10:29
Inference - Semi-supervised learning of unaries11:44
Inference – Example for mixture of trees model (1)12:13
Inference – Example for mixture of trees model (2)12:36
Inference – Example for mixture of trees model (3)12:38
Inference – Example for mixture of trees model (4)12:44
Road map (5)13:06
Results – Girl & Cheetah sequences from Segtrack13:08
Results - Quantitative evaluation on SegTrack13:55
Results – Bird-fall sequence15:03
Results – Parachute & Monkeydog sequences15:37
Summary15:50