Researchers at Pennsylvania State University have launched a new automatic picture indexing system that uses ‘machine intelligence’ to attach indexing terms to digital images, thereby making them more visible to Web search engines.

 

November 5, 2006: Researchers at Pennsylvania State University have launched a new automatic picture indexing system that uses ‘machine intelligence’ to attach indexing terms to digital images, thereby making them more visible to Web search engines.
The ALIPR (pronounced a-lip-er) system was launched on November 1, 2006, and is the brainchild of Professors Jia Li and James Z. Wang. They started their research in 1995 at Stanford University, where the Yahoo! and Google projects were started by their fellow schoolmates. However, Li and Wang have taken a different tack by developing computer systems to manage millions of images by the pixel content. ALIPER is simple to use; a browser interface allows photographers to upload an image from their own collection. The image is then analysed – a process that takes several minutes with large files – and the photographer is presented with a list of words that can be applied to the picture. Because ALIPR is still in a ‘learning’ stage, users are being encouraged to select the words that best apply to the picture and add any others that may have been omitted.
Photographers who use the Firefox browser can simply drag and drop an image from a Web page into the URL field provided or use the Ctrl+right mouse button on the image to call up a menu that includes “Copy Image Location” and directly paste the image into ALIPR. Flikr users can use the View Image function when the menu pops up, then copy the URL of the image from the browser’s URL field. ALIPR provides the following benefits to photographers:
1. It makes pictures more visible to the search engines and hence to Web surfers.
2. It can save professional photographers and photographers who sell images to stock libraries time when compiling image metadata.
3. It’s fun to use.
4. Uploaded pictures are featured on ALIPR site.
5. It allows users to participate in the development of ALIPR as it evolves.

ALIPR is Patent Pending. Version 1.0 has a vocabulary of 332 English words at the moment. It’s designed for use only with colour photographic images and cannot be used with black & white photos, manipulated images, cartoons, sketches or framed photos. Details can be found at http://www.alipr.com/about.html. You can try it out at http://www.alipr.com/cgi-bin/alipr_upload.cgi.