This illusion is one of several effects in which a rigid object is shown moving behind apertures in an occluding
surface. In this case incomplete information about the occluding surface causes the visual system to alternate between different figure-ground organisations, which correspond to four different percepts.
1) A black ellipse rotating behind a cross-shaped aperture.
2)
A black cross whose ends shrink and expand cyclically.
3) A stationary cross seen through a rotating elliptical aperture.
4) Two black bars oscillating in depth.
Mid-level motion processes serve to build representations of the shape of moving surfaces by integrating information across large image areas. In this case insufficient information is available to reach a definitive solution, so the visual system continuously cycles through alternative possible organisations, perhaps driven by moment to moment fluctuations in the responsiveness of different groups of neurons.
See:
Caplovitz, G. P., & Tse, P. U. (2006). The bar-cross-ellipse illusion: Alternating percepts of rigid and nonrigid motion based on contour ownership and trackable feature assignment. Perception, 35, 993-997.