Something that has fascinated me with the evolution of the web have been the ever changing methods in which users are able to describe content on the web. The idea behind describing content is a rather simple one. You can take an image, sound or video and use words to describe it.

Lately, there is becoming a more prevalent means of describing content that operates at a much lower level. It plays on the subtle differences between describing a final piece of content and describing the attributes that make up a final piece of content. So instead of describing a video as funny or as containing Patrick Swayze and Keanu Reeves,  It uses this idea of genealogy to generate a wide array of meaningful “genes”; things like format, quality, genre etc… Pandora uses this method to describe and relate music. This same idea is now being applied to imagery in the Glocal Project. Using Flickr and a series of custom scripts, the project is able to record image attributes such as color palette and luminosity as well as other more standard data like descriptive tags, location and more. This data is then used to create a unique signature for each image in the Glocal pool. A series of comparisons are done on two “parent” images to locate similar child images. These comparisons can be done over multiple generations to create a sort of lineage of imagery.

This could potentially evolve in to useful way to build and browse through large photographic archives.

Photos can be contributed to the project by uploading them to flickr and tagging them with “glocalproject” and joining the glocal group.