Tomographic Stringart (TS), as we call it, has its roots in a series of lectures about medical image reconstruction. During a Computed Tomography (CT) scan, rays of light travel through the body at different angles in straight lines and get absorbed to a greater or lesser extend based on the type of tissue in between. Light rays, straight lines, a thread fixed between two nails – the idea was born.

Since the nails are placed only at the border of our shapes, the contrast in the middle is simply build up by more strings crossing each other at darker places while less strings passing through brighter regions. For non trivial patterns, like faces, the challenge of finding a good sequence of pins to be connected by one single continuous string is fascinating. Like in real life, the generation of a TS consists of a sequence of choices, that is to say which pin to take next. Most of these choices are natural, while some need deeper consideration to make sense in the big picture.