Suggestions Needed

[socrtwo]

I want to set up imgseek server so that it matches uploaded images
with a subset and maybe eventually all the wikipedia/wikimedia
images. One can download a dump of all the images using this script:
http://meta.wikimedia.org/wiki/Wikix.

I will download the images to a Ubuntu computer. There may be 500 GB
worth of pictures, which is ok. It should take about 100 hours
supposedly. What I want to do is use image seek to allow users to
upload their images and then get possible matches from Wikipedia. I
don't want to store a mirror of all the images to my hosting company.
I think I want to store the URL's of the Wikimedia pages of images
matching user uploaded images and user when they click on a retrieved
image, would be taken to the appropriate Wikimedia page.

Does anybody have any suggestions? I'll have to replace the local
addresses with the web addresses before uploading the isk-daemon
server database up to my hoster.

[Maximilian Muenchow]

On Tue, Jul 07, 2009 at 08:48:57PM -0700, socrtwo wrote:
> upload their images and then get possible matches from Wikipedia. I
> don't want to store a mirror of all the images to my hosting company.
> I think I want to store the URL's of the Wikimedia pages of images
> matching user uploaded images and user when they click on a retrieved
> image, would be taken to the appropriate Wikimedia page.
>
> Does anybody have any suggestions? I'll have to replace the local
> addresses with the web addresses before uploading the isk-daemon
> server database up to my hoster.
>

Hello,

isk-deamon saves all information in its own database so the images could be
deleted right after they have been indexed.
About retriving the links of a image: Im not sure if isk-deamon can save that
information in its db, but in the worst case its still possible to store
that information in a separate db.