I’m generally drawn to design/architecture/art that makes dramatic use of negative space. Last night as I lay in bed, I began to wonder if a computer program could process images and detect the severity of an image’s whitespace.

Write a program (I would probably use Actionscript) that processes each pixel in an image, aggregating data about the average luminosity (or hue, or saturation, or even a combination of all three) for each row and/or column of pixels. From this data, the luminosity of a row/column could be treated as mass: that is, dark areas in a highly luminous image are ‘heavy’, or vice versa: light areas in a dark image are ‘heavy’.

Here’s a sketch to clarify the basic concept. The shaded object on the right gives weight to the right side of the canvas, shifting its perceived center of gravity to the right:

As for the potential application of such an algorithm, I’m not sure. Perhaps it could be used to find patterns among an artist or designer’s œuvre..

Mark Rothko

Cristiana Couceiro

Wolfgang Tillmans

Has anyone seen such a thing? What are some other possible applications of this concept?