Image Search and Annotation on a Very Large Scale
Using Deep Features and MI-File to search and annotate YFCC100M
(Yahoo Flickr Creative Commons 100M) dataset.

VEMIR
Research group

(97 millions images currently indexed)
Upload an image as query:
Chose an url as query:
Choose a random query.

Found 801953 images similar to your query in 1051 ms (Index access Time: 109 ms, Temporary result managemet time: 313 ms, Database access time: 629 ms)

Query:

Best guessed tags:
Relevance: 16.02, Guessed tags: paul
Relevance: 10.76, Guessed tags: cathedral
Relevance: 7.12, Guessed tags: london
Relevance: 6.20, Guessed tags: capitol
Relevance: 5.34, Guessed tags: dome
Relevance: 4.43, Guessed tags: build
Relevance: 4.40, Guessed tags: church
Search using images similar to your query or see more random images:


(Weight 0) Flickr->

(Weight 251) Flickr->

(Weight 263) Flickr->

(Weight 277) Flickr->

(Weight 285) Flickr->

(Weight 298) Flickr->

(Weight 308) Flickr->

(Weight 311) Flickr->

(Weight 316) Flickr->

(Weight 317) Flickr->

(Weight 323) Flickr->

(Weight 328) Flickr->

(Weight 333) Flickr->

(Weight 339) Flickr->

(Weight 341) Flickr->

(Weight 343) Flickr->

(Weight 344) Flickr->

(Weight 349) Flickr->

(Weight 349) Flickr->

(Weight 350) Flickr->
The above images are miniature versions of the original ones that come from the site www.flickr.com.
All rights are reserved to the author of the original image, which is directly available on the Flickr site through a hypertext link (surface linking).