0.25
0.5
0.75
1.25
1.5
1.75
2
Category-Specific Object Recognition and Segmentation Using a Skeletal Shape Model
Published on Dec 01, 20094152 Views
Related categories
Chapter list
Skeleton-search:Category-specific object segmentation/recognition using a skeletal shape model00:00
Category-specific object recognition00:10
giraffe (1)00:38
giraffe (2)00:46
giraffe (3)00:54
giraffe (4)01:02
Faces of Object Recognition01:06
Our goal01:42
Top-down approach01:54
Contributions02:34
1. How to represent shapes of an object category?03:08
Giraffes in Images03:09
Giraffe Shapes03:24
Giraffe Skeleton03:40
Shared Skeletal Topology03:52
Idea: Represent a shape using its skeleton03:56
Intrinsic Symmetry-based Shape Model04:04
Intrinsic Shape Model for Segmentation04:40
Fragment-Based Generative Model for Shape05:08
Reconstructing a Shape Fragment’s Boundary05:48
Generative Model06:04
2. How to measure support for a shape in an image?06:16
Cost function06:20
Cost of a shape fragment06:32
Oriented Chamfer Matching (OCM) (1)07:04
Oriented Chamfer Matching (OCM) (2)07:28
Drawbacks of OCM07:40
Improvement: Contour Chamfer Matching (CCM)08:20
How to search for the best supported shapes?09:16
Single Global Solution09:20
The need for multiple solutions09:48
Single-Pass Multiple Solution Using DP10:00
Experiments10:40
Dataset: ETHZ Shape Classes10:43
Detection / Segmentation -Giraffes11:06
Detection/Segmentation -Bottles11:26
Detection/Segmentation -Swans11:38
Detection/Segmentation -Applelogos 11:46
Detection/Segmentation -Mugs11:56
False Positives12:00
Object Detection Evaluation12:14
Object Detection Performance12:42
Evaluation: Segmentation Performance13:14
Performance –boundary coverage14:06
Performance –Boundary Precision14:46
Summary15:02
Thank you, Questions?15:32