en-de
en-es
en-fr
en-pt
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
What can the world tell us about an image?
Published on Dec 01, 200912438 Views
Related categories
Chapter list
What can we learn from a single image?00:00
What do we see?00:29
The Miserable Life of an Object Detector01:41
What the Detector Sees01:57
State-of-the-Art Pedestrian DetectionTrue02:07
Importance of Looking Globally02:51
Detail on Monet's painting03:27
Seeing less than you think… (1)04:33
Seeing less than you think… (2)05:29
Real Relationships are 3D06:17
Imaging Process06:57
Unsolvable Problem07:37
Ecological Optics07:53
Our World is Structured08:21
Understanding Scenes09:17
Support11:05
Size11:25
Interposition11:49
Position, Probability, Size11:53
+ illumination12:04
Our Goal: Scene Understanding12:24
Ecological Statistics13:30
Collaborators14:18
Scene Layout15:34
Learn from labeled data16:38
What cues to use?16:58
Weak Geometric Cues17:58
Need Good Spatial Support18:18
Image Segmentation19:30
Estimating surfaces from segments20:54
Labeling Segments21:22
Image Labeling21:58
Results22:18
No Hard Decisions22:45
Indoor Images23:29
Paintings23:33
Graphics application:Automatic Photo Pop-up23:41
Automatic Photo Pop-up24:30
Failures25:32
Occlusions are everywhere!25:52
Finding occlusions26:45
Occlusion Reasoning as Classification27:53
Object Size / Camera ViewpointWorld28:55
Camera viewpoint for LabelMe30:07
Helping Object Detection (1)32:27
Helping Object Detection (2)33:19
More Chickens, More Eggs…33:59
Best Guesses34:30
Putting it all together35:22
Some Results (1)35:31
Some Results (2)35:51
More Results36:03
Putting Objects in Perspective36:35
Illumination from a Single Image (1)36:55
Illumination from a Single Image (2)38:19
Illumination from a Single Image (3)39:07
Algorithm39:27
Weak cues40:43
Data-driven Sun Elevation Prior (1)43:27
Data-driven Sun Elevation Prior (2)44:19
Scene Semantics:Understanding the EntireScene44:44
Hays & Efros, SIGGRAPH‘07 (1)45:16
Hays & Efros, SIGGRAPH‘07 (2)45:32
Where does the knowledge come from?45:40
Scene Semantics!45:56
Images example46:31
Scene Completion Result46:54
Example, Scene Descriptor (1)47:01
Example, Scene Descriptor (2)47:09
Example, Scene Descriptor (3)47:12
Gist scene descriptor (1)47:32
Gist scene descriptor (2)47:56
Gist scene descriptor (3)48:08
... 200 total48:15
Gist scene descriptor (4)48:21
Graph cut + Poisson blending (missing picture examples) (1)48:29
Example, Scene Descriptor (4)48:48
Example, Scene Descriptor (5)48:56
Example, Scene Descriptor (6)48:59
Example, Scene Descriptor, ... 200 scene matches (1)49:00
Example, Scene Descriptor (7)49:12
Example, Scene Descriptor (8)49:40
Example, Scene Descriptor (9)49:43
Example, Scene Descriptor (10)49:47
Example, Scene Descriptor (11)49:49
Example, Scene Descriptor (12)49:53
Example, Scene Descriptor (13)49:58
Example, Scene Descriptor (14)50:02
Example, Scene Descriptor (15)50:03
Example, Scene Descriptor, ... 200 scene matches (2)50:12
Example, Scene Descriptor (16)50:31
Example, Scene Descriptor (17)51:07
Why does it work?51:11
Example, Scene Descriptor (18)51:22
10 nearest neighbors from acollection of 20,000 images (1)51:30
10 nearest neighbors from acollection of 20,000 images (2)51:54
Database of 70 Million 32x32 images52:18
The Big Picture52:46
im2gps (Hays & Efros, CVPR 2008)53:19
How much can an image tell about its geographic location?53:51
Image (and its geographic location), example 154:02
(Image and) its geographic location, example (1)54:08
Im2gps (1)54:32
Example Scene Matches55:16
Voting Scheme55:24
im2gps (2)55:44
Image (and its geographic location), example 256:30
(Image and) its geographic location, example (2)56:40
Image (and its geographic location), example 356:44
(Image and) its geographic location, example (3)57:01
Image (and its geographic location), example 457:25
(Image and) its geographic location, example (4)57:26
Image (and its geographic location), example 557:39
(Image and) its geographic location, example (5)57:41
Image (and its geographic location), example 658:02
(Image and) its geographic location, example (6)58:06
Data-driven categories58:22
Data-driven categories, geographic location58:45
Data-driven categories, Elevation gradient = 112 m / km*59:01
Elevation gradient magnitude ranking59:32
Figure 2, Global population density map59:42
Population density ranking59:46
Figure 4, Global land cover classification map59:49
Barren or sparsely populated59:59
Urban and built up01:00:00
Snow and Ice01:00:03
Savannah01:00:15
Water01:00:17
Conclusions01:00:19