Enter & Re-Exit takes the most image laden form of word, poetry, and juxtaposes how Google, which ranks image results according to its own notion of relevance (perhaps a pop culture in itself), views it with how each of us, as individuals, views it.

Its goal is to invite viewers to ask: What is the connection? This process will expand our interpretation and give us a view into another consistent, unfamiliar web of connections.

It achieves this by presenting the poem in two ways:

• As a recursive composition of images, where each image is made up of images with similar gestalt properties
• By coloring each phrase in a poem according to Google's synesthesia

Submission goal: electrofringe 2008 online submission

Enter & Re-Exit uses basic techniques to break an image into a graph of blobs, where each blob is a group of connected, similar looking pixels.

If you look at just the blob graph of most images online, you'll find an overwhelming amount of black and dark colors. Because we see the color in the darkness, not the darkness itself (Gestalt), another transform is needed to make the blob graph useful.

The blob graph is transformed into a perceptual graph using heuristics like:

• A region of color surrounded by a larger region of dissimilar color stands out
• Large areas of colors are less prominent
• Less frequently used color stand out
• Subtle variations in color amplify the texture, and perceptibility, of a region

Using the perceptual graph, each image is classified with its most prominent colors. A database that maps prominent colors to images is maintained, from which we construct the visualizations of each poem.