The intervals are defined by regions of the image that are too light or too dark – the edges of these regions define the edges of the intervals. Kim Asendorf’s code simply applies this technique twice: first vertically, then horizontally. The pixels in each interval are rearranged so as to sort them with respect to some property, say, lightness.The rows/columns of the image are split into ‘intervals’ (more on this later).They propose using it as a technique for unsupervised classification, to be applied to images from remote sensing satellites. Neat! The earliest reference to pixel sorting I could find on the internet is this paper ( PDF) by scientists from Iraq’s Scientific Research Council. The processing script was cryptic and not very hackable, and I felt like something more lightweight was needed, so I wrote my own version in python – more info on GitHub.
It was popularized (possibly invented) by artist Kim Asendorf (processing source code here). Pixel sorting is an interesting, glitchy effect which selectively orders the pixels in the rows/columns of an image. Pixel Sorting Introducing order in a disordered way can make for some great glitch art.
