An image is drawn onto a sheet of foam core. Depending on the sheet’s angle and location, the image morphs into different shapes and designs. These shapes aren’t arbritrary; they reveal a cross section of an otherwise invisible 3d model. The 3d model is a replica of the cardboard box – including the box’s contents.
This video is sketch of this idea. Ultimately, the goal would be to do this on the fly using a projector and motion capture. Imagine reading brain CAT Scans in a dynamic cross section renderer instead of looking at a series of slides.