Animals on the Edge Global warming and its Impact on Animals.
Animals on the Web
description
Transcript of Animals on the Web
![Page 1: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/1.jpg)
UCB
Compu
ter
Vision
Animals on the Web
Tamara L. Berg
CSE 595 Words & Pictures
![Page 2: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/2.jpg)
UCB
Compu
ter
Vision
I want to find lots of good pictures of monkeys…
What can I do?
![Page 3: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/3.jpg)
UCB
Compu
ter
Vision
Google Image Search -- monkey
Circa 2006
![Page 4: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/4.jpg)
UCB
Compu
ter
Vision
Google Image Search -- monkey
![Page 5: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/5.jpg)
UCB
Compu
ter
Vision
Google Image Search -- monkey
![Page 6: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/6.jpg)
UCB
Compu
ter
Vision
Google Image Search -- monkey
Words alone won’t work
![Page 7: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/7.jpg)
UCB
Compu
ter
Vision
Flickr Search - monkey
Even with humans doing the labeling, the data is extremely noisy -- context, polysemy, photo sets
Words alone still won’t work!
![Page 8: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/8.jpg)
UCB
Compu
ter
Vision
Our Results
![Page 9: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/9.jpg)
UCB
Compu
ter
Vision
General Approach
- Vision alone won’t solve the problem. - Text alone won’t solve the problem.
-> Combine the two!
![Page 10: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/10.jpg)
UCB
Compu
ter
Vision
Previous Work - Words & Pictures
Barnard et al,CVPR 2001
Clustering Art
Barnard et al, JMLR 2003
Labeling Regions
![Page 11: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/11.jpg)
UCB
Compu
ter
Vision
Animals on the Web
Extremely challenging visual categories.
Free text on web pages.
Take advantage of language advances.
Combine multiple visual and textual cues.
![Page 12: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/12.jpg)
UCB
Compu
ter
Vision
Goal:
Classify images depicting semantic categories of animals in a wide range of aspects, configurations and appearances. Images typically portray multiple species that differ in appearance.
![Page 13: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/13.jpg)
UCB
Compu
ter
Vision
Animals on the Web Outline:
Harvest pictures of animals from the web using Google Text Search.
Select visual exemplars using text based information.
Use visual and textual cues to extend to similar images.
![Page 14: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/14.jpg)
UCB
Compu
ter
Vision
Harvested Pictures
14,051 images for 10 animal categories.
12,886 additional images for monkey category using related monkey queries (primate, species, old world, science…)
![Page 15: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/15.jpg)
UCB
Compu
ter
Vision
Text Model
Latent Dirichlet Allocation (LDA) on the words in collected web pages to discover 10 latent topics for each category.
Each topic defines a distribution over words. Select the 50 most likely words for each topic.
1.) frog frogs water tree toad leopard green southern music king irish eggs folk princess river ball range eyes game species legs golden bullfrog session head spring book deep spotted de am free mouse information round poison yellow upon collection nature paper pond re lived center talk buy arrow common prince
Example Frog Topics:
2.) frog information january links common red transparent music king water hop tree pictures pond green people available book call press toad funny pottery toads section eggs bullet photo nature march movies commercial november re clear eyed survey link news boston list frogs bull sites butterfly court legs type dot blue
![Page 16: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/16.jpg)
UCB
Compu
ter
Vision
Animals on the Web Outline:
Harvest pictures of animals from the web using Google Text Search.
Select visual exemplars using text based information.
Use vision and text cues to extend to similar images.
![Page 17: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/17.jpg)
UCB
Compu
ter
Vision
Select ExemplarsRank images according to whether they have these likely words
near the image in the associated page (word score)
Select up to 30 images per topic as exemplars.
2.) frog information january links common red transparent music king water hop tree pictures pond green people available book call press ...
1.) frog frogs water tree toad leopard green southern music king irish eggs folk princess river ball range eyes game species legs golden bullfrog session head ...
![Page 18: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/18.jpg)
UCB
Compu
ter
Vision
SensesThere are multiple senses of a category within
the Google search results.
Ask the user to identify which of the 10 topics are relevant to their search. Merge.
Optional second step of supervision – ask user to mark erroneously labeled exemplars.
![Page 19: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/19.jpg)
UCB
Compu
ter
Vision
Image Model
Match Pictures of a category
![Page 20: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/20.jpg)
UCB
Compu
ter
Vision
Geometric Blur Shape Feature
Sparse Signal Geometric Blur
(A.) Berg & Malik ‘01
Captures local shape, but allows for some deformation. Robust to differences in intra category object shape.
Used in current best object recognition systems Zhang et al, CVPR 2006Frome et al, NIPS 2006
![Page 21: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/21.jpg)
UCB
Compu
ter
Vision
Image Model (cont.)Color Features: Histogram of what colors appear in the image
Texture Features: Histograms of 16 filters
* =
![Page 22: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/22.jpg)
UCB
Compu
ter
Vision
Animals on the Web Outline:
Harvest pictures of animals from the web using Google Text Search.
Select visual exemplars using text based information.
Use vision and text cues to extend to similar images.
![Page 23: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/23.jpg)
UCB
Compu
ter
Vision
++
++
+
++
+ ++
++
+
+
++
++
***
**
*
* *
**
*
**
**
* **
*
+ *+
+
+
+
+
+ +
+*
*
?
*
+++
*
?
+
+
Scoring ImagesRelevant Features
Irrelevant Features
Relevant Exemplar
For each query feature apply a 1-nearest neighbor classifier. Sum votes for relevant class. Normalize.
Combine 4 cue scores (word, shape, color, texture) using a linear combination.
Query
*
Irrelevant Exemplar
![Page 24: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/24.jpg)
UCB
Compu
ter
Vision
Classification ComparisonW
ord
sW
ord
s +
Pic
ture
![Page 25: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/25.jpg)
UCB
Compu
ter
Vision
Cue Combination:Monkey
![Page 26: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/26.jpg)
UCB
Compu
ter
Vision
Cue Combination:
FrogGiraffe
![Page 27: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/27.jpg)
UCB
Compu
ter
Vision
Re-ranking Precision
Classification Performance
![Page 28: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/28.jpg)
UCB
Compu
ter
Vision
Re-ranking Precision
Monkey Category
Classification Performance
Monkey
![Page 29: Animals on the Web](https://reader035.fdocuments.net/reader035/viewer/2022070405/56814037550346895daba2a2/html5/thumbnails/29.jpg)
UCB
Compu
ter
Vision
Ranked Results:
http://tamaraberg.com/google/animals/index.html