Toward a digitally reconfigurable optical cloak

May 23, 2016 // By Julien Happich
In their recent Optica paper “digital integral cloak”, researchers from the University of Rochester combine a scanning camera on one side and an LCD display overlaid with an array of micro-lenses on the other side of the object to be concealed, together with ray-tracing algorithms running in between.

The proof-of-concept demonstration tricks the eyes into seeing what's behind the cloaked object as if the light rays were directly passing through the "invisible" object. The cloak is reconfigurable in the sense that if the background changes, the output can be updated through another scan of the background area by the camera, with ray-tracing algorithms put into play to compute the right display rendering on the LCD so each micro-lens redirects the right colours in the right directions (the maths have to take into account the depth of the concealed zone).


Fig. 1: Cross section of a digital integral cloak showing two parallel 2D surfaces, with a few sample rays. The input “surface” (lens array and plate) captures input light rays. The output surface displays rays as if they passed through ambient space only (dashed lines). Superpixels, placed at the focusing plane of a lenslet, collect rays with the same position as the lens. These rays are then spatially separated into pixels, such that one ray angle (or “view”) maps to one pixel. Display (output) is the reverse of the detection scheme.

The thin, parallel semi-cylindrical micro-lenses then recreate multiple images of the background, hence perfecting the illusion regardless of the viewer's position (a considerable improvement over a flat 2D rendering which would equate to intercalating a poster of the background).

Because the proof-of-concept experiment only used a camera mounted onto a rail for scanning the background, it took PhD student Joseph Choi and his advisor Professor of Physics John Howell several minutes to scan, process and update the image on the screen for every change in background.

But in the future, Choi envisages that one side of the object to be concealed could be covered with a lenticular array of optical sensors, and a conformable lenticular display onto its other side showing the right pixels in all directions so as to show the background as if the object weren't there. A fixed setup would only work as long as the object's shape wouldn't change, the geometry and spatial location of the optical sensors being taken into consideration in the mathematical model to compute which pixels to feed on the output display.