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 1280536 images similar to your query in 1266 ms (Index access Time: 95 ms, Temporary result managemet time: 546 ms, Database access time: 625 ms)

Query:

Best guessed tags:
Relevance: 7.32, Guessed tags: track
Relevance: 6.37, Guessed tags: skyscraper
Relevance: 6.36, Guessed tags: tower
Relevance: 4.87, Guessed tags: railway
Relevance: 4.83, Guessed tags: why
Relevance: 4.78, Guessed tags: build
Relevance: 4.10, Guessed tags: train
Relevance: 3.96, Guessed tags: empire state
Search using images similar to your query or see more random images:


(Weight 0) Flickr->

(Weight 586) Flickr->

(Weight 608) Flickr->

(Weight 611) Flickr->

(Weight 623) Flickr->

(Weight 631) Flickr->

(Weight 653) Flickr->

(Weight 661) Flickr->

(Weight 671) Flickr->

(Weight 681) Flickr->

(Weight 684) Flickr->

(Weight 687) Flickr->

(Weight 688) Flickr->

(Weight 689) Flickr->

(Weight 692) Flickr->

(Weight 698) Flickr->

(Weight 699) Flickr->

(Weight 700) Flickr->

(Weight 702) Flickr->

(Weight 703) 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).