en
0.25
0.5
0.75
1.25
1.5
1.75
2
Using Richer Models for Articulated Pose Estimation of Footballers
Published on Oct 09, 20123901 Views
We present a fully automatic procedure for reconstructing the pose of a person in 3D from images taken from multiple views. We demonstrate a novel approach for learning more complex models using SVM
Related categories
Chapter list
Using Richer Models for Articulated Pose Estimation of Footballers00:00
What?00:11
Why?00:35
How?00:50
Ecient Pose Estimation01:32
Flexible mixture of parts, Ramanan [2011]02:18
FMP on football data - 103:24
FMP on football data - 203:44
FMP on football data - 303:52
N-Best con gurations04:02
Rank of the best con guration04:36
Using a richer model05:13
Model deformation between unconnected parts06:00
Penalize Unexplained Pixels: FG/BG Modelling - 106:17
Penalize Unexplained Pixels: FG/BG Modelling - 206:53
Avoid Double Counting: an Exclusion Principle - 108:07
Avoid Double Counting: an Exclusion Principle - 208:50
Avoid Double Counting: an Exclusion Principle - 309:11
Reranking - 109:23
Reranking - 210:15
SVM Rank vs Linear SVM10:49
Dataset11:19
Results - 111:30
Results - 212:01
Results - 312:25
Results - 412:52
Results - 513:12
Results - 613:33
3D Reconstruction - 113:41
3D Reconstruction - 213:52
3D Reconstruction - 314:14
3D Reconstruction - 414:30
3D Reconstruction - 514:44
Results - 714:54
Results - 815:13