As opposed to seeding from randomness - this series started from my own abstract intuition. Vapid, reactive, and blue. Each piece began with large strokes of ink swashed around to find the shape, then filled out with a fountain pen. Later the image was segmented into an 11 x 15 matrix (165) squares. Many of these squares were then treated like their own small canvas. The intention of this segmentation was built around writing a program which would analyze and identify each of the smaller cubes.

Object Detection, a software which has many applications -advertising and marketing being one that is pervasive today- is used to classify objects. Rendering them semantically understood, feeding the predicatble, or making the environment more easily predicted. By using this software as an abstracting tool I hope to subvert its purpose, and take an action that is antithetical to the status quo.

Using this as a technique to play a canvas like instrument further abstracts the piece, the notes and pace of the synthesizers are derived from the confidence the machine has in each object - how certain it is that it's actually seeing a shoe or a pillow. Through this you can begin to hear the rhythm of its assumptions, like a purring animal investigating a rotting log.