The cellular neural/nonlinear network (CNN) has become a useful tool for image and signal processing, robotic and biological visions, and higher brain functions. visual illusion is defined as any fallacious perception of reality. In 1997, Chua proposed a CNN template named face-vase illusion (FVI) CNN. It can simulate some visual illusions. In this paper, we figure out the local rules of the FVI CNN. A theorem has been set up which proves that the FVI CNN does satisfy the local rules. An interesting function of the FVI CNN is that it can also detect the connectivity in an image. Some numerical simulations are given.