Deep Learning for visually searching and annotating images on a very large scale

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)

Choose a random query.

Found 655556 images similar to your query in 521 ms (Index access Time: 93 ms, Temporary result managemet time: 141 ms, Database access time: 287 ms)

Query:

Best guessed tags:
Relevance: 9.30, Guessed tags: draw
Relevance: 5.46, Guessed tags: sketch
Relevance: 4.67, Guessed tags: cartoon
Relevance: 3.91, Guessed tags: then

Search using images similar to your query or see more random images:

(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 0) Flickr->
(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 0) Flickr->
(Weight 0) Flickr->	(Weight 0) Flickr->	(Weight 215) Flickr->	(Weight 215) Flickr->	(Weight 215) Flickr->
(Weight 215) Flickr->	(Weight 215) Flickr->	(Weight 215) Flickr->	(Weight 215) Flickr->	(Weight 215) 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).