iMMem

On this page you'll find some examples of iMMem technology on some internet still image collections.

For a demo of iMMem applications on video processing, see the demo page at TvTrac.

All the image collections works the same way, just paste the url of the image you want to search on the 'imageUrl' field and press the 'search' button.

The server side will download the requested image, compute his aspect information and launch the search against the selected collection, if the requested image is not accesible, an error page will be displayed.

In the results page, clicking in any of the returned images, will launch another search this time from this result image. In most collections there's also the original url of the result image at indexing time at your disposal.

It's possible that some resulting images does not exist anymore on the stored url, in those cases the image rectangle will be void. It's important to note that iMMem once indexed, doesn't need to store the graphical information, because it is not used at all in the search process.

Big Collections

200000 random images from Google+Yahoo indexed 2007-Oct.

Institutional Art Collections.

Joconde Collection from the Catalogue of French Museums from French Culture Ministry, 11000 art images ofartists named from A to D.

Art Collections.

1700 clasical pictures and lythographies from WebMuseum Paris.
3000 clasical paintings from the  CGFA site