en
0.25
0.5
0.75
1.25
1.5
1.75
2
Transfer Learning by Ranking for Weakly Supervised Object Annotation
Published on 2012-10-093700 Views
Most existing approaches to training object detectors rely on fully supervised learning, which requires the tedious manual annotation of object location in a training set. Recently there has been an
Related categories
Presentation
Transfer Learning by Ranking for Weakly Supervised Object Annotation00:00
Object Detection00:19
Existing Work for Object Detection00:58
Weakly Supervised Learning02:33
Existing Work for WSL03:23
Saliency [B.Alexe, CVPR 2010]03:38
Inter-class [T.Deselaers ECCV’10 , P.Siva ICCV’11] (1)04:27
Inter-class [T.Deselaers ECCV’10 , P.Siva ICCV’11] (2)06:12
Motivation07:14
Our approach: Transfer Laearning (TL)08:06
Transferrable Knowledge09:39
Annotating12:31
Ground Truth Approximation (1)12:45
Ground Truth Approximation (2)13:04
Proposed Approach13:24
Experiments (1)13:45
Experiments (2)14:13
Experiments (3)15:53
Experiments (4)16:04
Summary16:20
Thank you!16:56