Action Recognition with Exemplar Based 2.5D Graph Matching thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Action Recognition with Exemplar Based 2.5D Graph Matching

Published on Nov 12, 20126893 Views

This paper deals with recognizing human actions in still images. We make two key contributions. (1) We propose a novel, 2.5D representation of action images that considers both viewindependent pose in

Related categories

Chapter list

Action Recognition with Exemplar Based 2.5D Graph Matching00:00
Action Recognition in Still Images (1)00:07
Action Recognition in Still Images (2)00:18
Action Recognition in Still Images (3)00:26
Action Recognition in Still Images (4)00:26
Action Recognition in Still Images (5)00:33
Action Recognition in Still Images (6)01:07
Action Recognition in Still Images (7)01:14
Action Recognition in Still Images (8)01:22
Outline (1)01:36
Outline (2)01:49
2.5D Representation of Actions (1)01:50
2.5D Representation of Actions (2)01:54
2.5D Representation of Actions (3)02:04
2.5D Representation of Actions (4)02:12
2.5D Representation of Actions (5)02:22
Estimating 2D Key Points02:52
Converting 2D Key Points to 3D03:19
Outline (3)04:23
Matching Two 2.5D Graphs (1)04:29
Matching Two 2.5D Graphs (2)04:57
Exemplar Based Action Recognition (1)05:11
Exemplar Based Action Recognition (2)05:22
Exemplar Based Action Recognition (3)05:28
Exemplar Based Action Recognition (4)05:35
Finding Dominating Images (1)05:45
Finding Dominating Images (2)06:12
Finding Dominating Images (3)06:19
Finding Dominating Images (4)06:27
Outline (4)06:38
PPMI: Dataset06:41
PPMI: Results (1)06:53
PPMI: Results (2)07:09
PPMI: Results (3)07:36
PPMI: Results (4)07:52
36 PPMI: Dominating Images (1)08:41
36 PPMI: Dominating Images (2)08:57
PASCAL 2011: Dataset09:07
PASCAL 2011: Pose Estimation09:10
PASCAL 2011: Results (1)09:34
PASCAL 2011: Results (2)09:47
PASCAL 2011: Results (3)10:06
Conclusion10:29
Acknowledgement10:46