0.25
0.5
0.75
1.25
1.5
1.75
2
Latent SVMs for Human Detection with a Locally Affine Deformation Field
Published on Oct 09, 20125071 Views
Methods for human detection and localization typically use histograms of gradients (HOG) and work well for aligned data with low variance. For methods based on HOG despite the fact the higher resolu
Related categories
Chapter list
Latent SVMs for Human Detection with a Locally Affine Deformation Field00:00
Human Detection (1)00:06
Human Detection (2)00:12
HOG Detector (1)00:17
HOG Detector (2)00:34
HOG Detector (3)00:37
HOG Detector (4)00:39
HOG Detector (5)00:42
HOG Detector (6)00:47
HOG Detector (7)00:50
HOG Detector (8)00:54
Deformable Part-based Model (1)01:08
Deformable Part-based Model (2)01:59
Comparison with other approaches02:27
HOG Detector (9)02:44
HOG Detector (10)02:58
Detector with Deformation Field (1)03:06
Detector with Deformation Field (2)03:14
Detector with Deformation Field (3)03:31
Detector with Deformation Field (4)03:50
Detector with Deformation Field (5)04:10
Detector with Deformation Field (6)04:35
Detector with Deformation Field (7)05:15
Locally Affine Deformation Field (1)05:41
Locally Affine Deformation Field (2)05:59
Locally Affine Deformation Field (3)06:05
Locally Affine Deformation Field (4)06:09
Locally Affine Deformation Field (5)06:12
Locally Affine Deformation Field (6)06:15
Locally Affine Deformation Field (7)06:23
Locally Affine Deformation Field (8)06:27
Locally Affine Deformation Field (9)06:40
Optimisation (1)07:05
Optimisation (2)07:18
Optimisation (3)07:27
Optimisation (4)07:53
Optimisation (5)08:03
Optimisation (6)08:39
Optimisation (7)08:48
Optimisation (8)09:15
Optimisation (9)09:38
Optimisation (10)09:53
Optimisation (11)10:21
Optimisation (12)10:32
Optimisation (13)11:34
Learning multiple poses / viewpoints (1)11:49
Learning multiple poses / viewpoints (2)12:15
Experiments12:27
Clustering of training samples13:01
Qualitative results (1)13:50
Qualitative results (2)14:07
Qualitative results (3)14:13
Qualitative results (4)14:21
Conclusion and Further Work15:16
Thank you15:50